https://www.theatlantic.com/science/archive/2018/06/its-possible-to-reverse-climate-change-suggests-major-new-study/562289/

https://www.theatlantic.com/technology/archive/2018/05/machine-learning-is-stuck-on-asking-why/560675/

nautil.us/blog/is-facebook-really-scarier-than-google

https://qz.com/1204395/self-driving-cars-trolley-problem-philosophers-are-building-ethical-algorithms-to-solve-the-problem/

Explainability

https://www.reuters.com/article/us-france-economy-trade/france-to-vet-takeovers-of-firms-in-data-and-artificial-intelligence-idUSKBN1F118H

There is significant effort invested right now in figuring out where this AI journey is going to take us. That’s a big question, realising how generations to come will be impacted. One of our favourite quotes here at II is ‘predictions are hard, especially when they’re in the future’. So we can safely say that most forecasts will miss. 

For clarity, when we think about AI we divide it into two segments. The first dates back to the 80s, which we can refer to as expert systems. It’s good stuff, but it’s the latest applications that are most exciting and impactful. In this context AI is code for ‘machine learning’. That’s the special sauce we’re fixated upon. 

Not unusually for tech, the hype is in any case way ahead of the reality. 

The AI market still needs to be primed. Dr Mike Lynch, whom runs Invoke Capital, reviews a large number of AI technologies. Under the marketing veneer he states that 90% contain no machine learning capability. Of the remaining 10%, 5% have not been market tested. This is not surprising. The Cloud shift demonstrated the same approach. Some providers, both longstanding and brand new, struggled to deliver against their messaging. And 20-years elapsed from the first conversations on utility computing to Cloud being mainstream.

We think AI is a more significant situation. It’s more sophisticated than Cloud for a start. The New York Times states that there are fewer than 10,000 people capable of AI research. That’s not many, and as things heat up it feels progressively less. Scarcity may be further exacerbated as tech organisations persuade academic AI researchers to leave their universities.

We still believe in watching out for the ‘next littlest thing’. Startups are able to directly address their targeted use case. Despite a scarcity of skills, and the cost of those that are available, that puts them at an advantage. 

Hopefully the queue of folks at Google’s Advanced Solutions Lab won’t be too long. We’ll look forward to seeing you there. 

Link: https://www.nytimes.com/2017/10/22/technology/artificial-intelligence-experts-salaries.html

At Infinite Insight we work with organisations that are large and complex, or undertaking substantial digital change, or both! Here’s a snapshot of our thoughts on the source of innovation. 

It feels that 2013 was a long time ago, but that is when we first considered incorporating data science into standard business practice. That was a combination of the surfacing of academic research in AI combined with the powerful force that is the technologists search for the next big thing.

But now we’re mainstream, whether it’s our friends at Satalia reorganising how Tesco delivers to our homes or NotCo’s algorithms deciding how vegan mayo should taste. At least mainstream in the sense that sophisticated AI is in use, but we have to carefully dissect a service to see it’s there. 

Some of the more adventurous promises have yet to materialise. I can’t talk with a bot whom processes my mortgage. And my local authority still prefers that I send them completed forms. When I interact with one of my many digital devices the interface is still largely dumb. What’s going on? 

If you’ve worked in any organisation and attempted to make a significant change then you probably know. It’s really hard, even when rational and readily justifiable. If it requires assembling a crack team of data scientists and allocating supporting resources it gets more difficult. And if that output might be challenging and maybe risky then more difficult still. 

And then there’s skills. Google’s Machine Learning Advanced Solutions Lab is a move to the IBM model of tech plus people. They have to assist us in our ambitions when combined with their platforms. Otherwise the vision will always be just that. Investing in professional services highlights the lack of scalability in AI.

Here at II we believe that smaller organisations such as Satalia and NotCo will lead the way. The challenges in transitioning existing processes to AI may be too great – why not buy-in? That’s why we continuously scour the market, not for the next big thing, but for the next littlest thing.

https://www.theguardian.com/science/2017/nov/01/cant-compete-universities-losing-best-ai-scientists