I’ve been reading with interest someone in Australia pulled out a letter stashed in a wall in 1995, that turned out to have some eerily prophetic guesses as to the future.
Now’s my turn. I promise not to be too negative. I’ll break that, but my intentions are not bad.
First up – Artificial Intelligence. Some of the work I do is in Machine Learning, and I am reasonably confident when I say that we’re a long way off General Artificial Intelligence. Just today we had Glenn Gore, chief architect for Amazon Web Services come and speak to a group of us, and the questions he took at the end mainly revolved around AI/ML, its applications and future. He gave a quick breakdown of all the companies who thought they were working on ML – 90% are working on clever spreadsheets – analytics. Clever, but not real ML. Of the remaining 10%, 9% of those were working on ways of adapting ML to simple use cases within their industry. I fall squarely into that category – we’re moving things out of a simple engineering environment and into a data science environment – and learning about feedback loops and reinforcement and how we can improve our system in this way. It’s very early stages for a lot of this stuff. The remaining 1% was deep learning, where the machine was developing new ideas, and basically this is the preserve of research institutions and places like Google. I only make mention of the stuff I am working on to highlight the simplicity of what we are doing – the use cases are so simple that the whole thing can be done faster and better using a logic engine and a set of correlations. No feedback required. The challenge we have is when a machine makes a decision, we need to be able to tell if it did a good job or not. Possible if your job was to sit there and monitor this thing, but not so easy if this is a minuscule part of the many things you do in your day.
We keep well abreast of AI/ML, and I both agree and disagree with Elon Musk’s statement about regulation of AI – yes, we should, but basically it’s impossible. The main drivers behind AI are state actors with nationalistic imperatives. The following talk is one of DARPA’s director of information innovation – this is the US military’s advanced research lab – the talk is both understandable and frightening if you consider the implications of the very last piece on contextual learning.
You can imagine that China has its own lab, along with other major nations. Now talk to me about regulation. End of. I would give AI about a 15% chance of a major breakthrough that changes everything, otherwise incremental achievements will finally get us to our goal, and given that neural networks have been around since the 80’s, we have a long way to go there.
Secondly – Resource Depletion. I think we have our finger in the dyke on this currently – we’re holding back little issues, in the knowledge that they will snowball over time and become critical. Just take fish as an example.
While I think this is a bit alarmist, this is hardly the point. The point is that we have this out of sight out of mind approach to a very rapidly depleting resource. Documentaries like “End of the Line” paint the decline out in graphic detail, along with the incredible amount of unwanted (dead) fish we just toss back. We have a 20th century attitude to a very 21st century problem, and the longer we leave these things unaddressed the more parts of any available solution disappear. Governments don’t want to deal with it even though the science seems reasonably clear.
Before too long I believe we will see one of the first big shocks – a resource that we have depended on drying up in very short order. Candidates are:
- Freshwater on a regional level. This is a global problem, but it will have regional effects first – we are already seeing this in places – sinkholes, aquifer depletion, water tables going down etc. There hasn’t been a city run dry yet, but some of China’s northern cities are running perilously close. Beijing’s water table has dropped 300 m since the 70’s:
- Phosphorous. You know, the fertiliser. We have around 50-100 years worth remaining.
- Fossil Fuels. You all know about this one. Depending on whether electrics and renewables kick in and to what degree, even if you have an electric car it still needs to draw its power from somewhere, and fossil fuels are still a massive part of the energy production mix.
You could go on – there’s lots. Point is more and more (population is expected to top 10 billion in this century) of us puts more and more strain on resources. As it is we’re heading towards zero on a lot – and I don’t subscribe at all to the idea that we’ll figure this one out like we did for food production in the 1800’s and all of the other obstacles we have overcome. We’re dealing with hard limits here – not just our abilities to adapt. This will becoming a sharply growing problem over the next 3 decades.
Third. Mental health problems, and societal dysfunction. The first two were obvious and very current in the media. This next one is something that I think is just starting to come into focus now. It is talked about in a range of ways, but none that I think really highlight the overall risk to society in a way that encompasses the problem.
From snowflake syndrome to social anxiety to the perception (and I do make the point of saying perception – because who knows…) of reduced social mobility, Governments washing their hands of social responsibility, the most unbelievable things happening in the US with regards to healthcare (to highlight the point of how much government is winding back its support), substance abuse epidemics etc.
I made another post some time ago about Universe 25, the mouse universe that tore itself apart because of abundance of everything, except it seems space. To me, space is absolutely a perceptual concept. Sure it has it’s hard edges – crowded stations, full lunch canteens – etc – these aren’t the problem. The problem as I see is that there isn’t the ability to disconnect anymore. It’s the very perception of being hemmed in by society, that you HAVE to be a part of this, don’t miss out, and in the midst of it you have some very bad actors, people are arbitrarily attacked for being in the wrong place at the wrong time, the rise of terrorist events, and the dislocation and dissociation of self – you can see above it’s little slices of problems, all gnawing at us from different angles.
I see this problem steadily continuing to deteriorate. And I don’t see any solution to it either, except for some major depopulation event, after which we can start to build up again. Already the natural population growth in 1st world countries is generally below the 2.1 children natural replacement rate. The population however is still expanding rapidly. This is being fueled by comparatively disadvantaged countries, with much lower standards of education – and we need these people to fill roles in our society. If you take Japan as an example of a country which restricts immigration to an incredible degree, there are currently on average 2 job openings for every applicant.
My gut feeling is that to all of these problems, it’s only after a massive shakeup – in whatever form that takes, that these problems become so overwhelmed by a much more basic issue – survival, that it takes a long time for the distortions to start appearing again. Case in point here is WW2 until now. After the war there were a number of fairly revolutionary things happen – the new deal (which was well underway by the time, but it was very popular and continued for decades afterwards), the welfare state, the Marshall Plan etc – all growth and support plans that encouraged social mobility, education, nation building etc. Growth was explosive, national debts were kept in check as a result of this growth, and many of the problems seemed fairly manageable. Contrast that to now – in the US, a growth rate of 2%, government debt to GDP of 106% and they want to talk about tax cuts and axing of medicare.
Mental health is an unseen problem – we’re not talking schizophrenia, dementia or anything obvious, we’re talking about our natural ability to deal with ever increasing stresses, every decreasing (it seems) empathy, and extrapolating from this what might be the long term consequences.
I dont think I am pushing things too far by saying that I feel this is the biggest underlying threat that we have. It will manifest itself in some major catastrophe, but this present and growing instability is one of the central causes.
You read what some people think on this subject – that we are going to back to the natural survival of the fittest mode of life where the strongest weed out the weaker and less deserving. Except when you have the ability to destroy the planet now, I’m not so sure that this way of thinking necessarily ensures any long term chances of survival. Over the next 30 years I think this idea will be put to the challenge like never before.