How open will the AI future really be?
Earlier this week, sitting in a café outside the Institut Suédois in Paris, an old friend from my art school days, visiting to work on a research project, mentioned something I had not heard about in a long time: the forgotten history of hypertext. Memories of a more open and optimistic, perhaps even naive, time flooded back.
Leaving aside the centralisation, or ‘siloisation’ of the Internet, what came to mind was how technologies and the companies who promote them rise and fall.
In the first decade of personal computing, a few companies came to dominate the entire sector. Pioneers such as Commodore and Tandy-Radio Shack clung on with remarkable tenacity, but such was the power of standardisation – and possibly branding – that a mere four giants bestrode the stage.
IBM was the dominant hardware manufacturer until it buckled in the face of an onslaught of cheap clones, but until that point it was the company that set the standard, and that standard was PCs with Intel CPUs and a Microsoft operating system.
As a result Intel, admittedly a pioneer, rapidly became the preeminent processor powerhouse. Its architecture set the standard to such a degree that it was widely cloned. Microsoft, meanwhile, crushed all in its path including its direct competitor in the operating system market Digital Research and countless application developers. Apple, meanwhile, was a kind of mini-giant, off in a corner serving a niche, but an embarrassingly profitable one that spared it the fate of its peers.
This all came to a head in the 1990s with attempts by the US courts to break up Microsoft’s monopoly. But then the world changed. What mattered now was the Internet, not PCs.
Today, Microsoft remains a behemoth but operating systems matter less than ever. Instead, the company’s profit comes from its grip on enterprise software, something it is clearly hoping to parlay into a regular income stream by converting users to the cloud, complete with its delicious monthly fees. Certainly, the company is far from the monopolistic threat it was in the 1990s, significantly trailing its cloud competitor Amazon.
Meanwhile Apple has transformed into a consumer electronics manufacturer and is greedily hunting down recurring subscription revenue. IBM no longer manufactures low- or mid-range hardware at all, instead concentrating on mainframes, high-end servers and services, with a side helping of quantum computing research that it hopes will be the next big thing. Intel remains massive but has been battered by competition from the likes of AMD and Nvidia and, worse still, missed the boat on mobile, effectively ceding the market to Arm licensees.
The next revolution
The lesson is, of course, that there is no point in re-fighting the last war. The Internet has come to be dominated by a different few companies: Google, Amazon and Facebook. Behind the scenes there are plenty of others engaged in networking or cloud, or overseeing content delivery networks (CDNs). however, as far as the public goes, the collective endeavours of this troika are very nearly the entirety of the Internet.
Will the same thing happen with AI? Now that we have got Web3 and blockchain out of our systems, thoughts are turning to the revolution that could unseat today’s dominant Internet businesses, and that thing might, or might not, be AI.
Webio, an Irish-based conversational AI provider that also operated in San Francisco and in Britain, certainly thinks there is an opportunity. The company today announced it has secured €2.5 million in new investment.
But there is more to this than ‘Irish-company-done-good’. AI is nothing new but there is no denying that since OpenAI unleashed ChatGPT on the world last year that it has become clear that it is AI that is the next big thing in technology, rather than the previous candidates: imaginary money or JPEGs of monkeys.
Since then, Microsoft has stuffed $10 billion into OpenAI and integrated its technology into all manner of products and services. Google has ramped-up the deployment of the AI technology we all knew it was busy developing. What is less clear is whether or not this will see them consolidate the market power of the US-based data scraping giants or if the technology will break open the market and allow others to challenge them.
Speaking to TechCentral.ie Paul Sweeney, Webio’s co-founder and chief strategy officer, noted that investors are betting on the plumbing that keeps the AI data flowing: CPU and GPU companies and infrastructure providers.
“Those lower down the stack are guaranteed to make money,” he said.
Start-ups and other pretenders to the crown will need to focus, he added, offering more than just a sprinkling of AI hype.
“New small companies can definitely eat the market share of much larger competitors, but have to find new propositions. Just doing what they did [but] using AI misses the point,” he said.
One risk with AI is that it will act to consolidate the grip of the likes of Google, Amazon and Facebook, as well as Microsoft. But with open source language models out in the wild, there is also the chance that, once again, everything could change.
Subscribers 0
Fans 0
Followers 0
Followers