In The Conversation

Neural nets are typically trained by “supervised learning”. So they’re presented with many examples of an input and the desired output, and then gradually the connection weights are adjusted until the network “learns” to produce the desired output.

To learn a language task, a neural net may be presented with a sentence one word at a time, and will slowly learns to predict the next word in the sequence.

This is very different from how humans typically learn. Most human learning is “unsupervised”, which means we’re not explicitly told what the “right” response is for a given stimulus. We have to work this out ourselves. 

Now is not the time for anyone to announce that their city will become “bigger and better”. Cities don’t have to get bigger to evolve, and sooner or later all will have to reckon with the concept of degrowth.

Australia must become less reliant on imports of skilled workers, students, tourists and materials. We can make better use of local resources and produce much more of what we need here.

Australian cities have very good bones. They have amazing cultural scenes. Their biomedical capabilities are among the world’s best. Our education sector remains eminently exportable online and via existing overseas campuses. The manufacturing sector still has a base to build on and provide many more of the products Australians need. And our renewable energy capacity is unlimited.

We can support our local hospitality and cultural venues better, and increase intercity and interstate patronage. We can invest in research and development and maintain wealth through innovation and production, rather than the eternal consumption of land.

The electric transport revolution is a great chance to rethink how we move through our cities – and whether we even need a car at all.

Cars, after all, often have only one occupant. You’re expending a lot of energy to transport yourself.

By contrast, electric mopeds and bikes use a lot less energy to transport one or two people. They’re also a lot cheaper to buy and run than electric cars.

If you commute on an e-bike 20km a day, five days a week, your charging cost would be about $20 – annually.

In 1990s Berlin, baugruppen (building groups) came to the fore in response to the German city’s housing crisis. Building groups are DIY collectives of future resident-owners who come together to develop their new homes. Households become producers rather than consumers, so they save on the developer’s profit margin and have more control over building design and quality.

At its peak, about 17% of new homes in Berlin were baugruppen projects. By 2017 more than 600 projects had been completed. The current master plan for redeveloping Berlin’s former Tegel airport calls for baugruppen to produce 2,000 homes – 40% of the project’s housing.

The success of baugruppen has inspired building groups in Australia. Data from one development and advisory service that assists building group members show members have on average saved around 10% on their new home building costs since 2010.

As well, they save on transfer taxes/stamp duties and mortgage interest payments. So in Victoria, for example, total savings could be as much as 16.5% on a A$1 million house.

In the years following the 2008 global financial crisis, the “big three” housebuilders that dominate the new-build market in Britain have been able to increase their profits without significantly increasing the number of homes they build. This has happened despite political pressure to increase UK housing supply.

They were able to do this, we argue, because they have built up significant structural power: they can use their control of housing land and housebuilding to secure state support for initiatives that benefit their shareholders by pushing up their share prices and profitability.

[…] 

We argue this state support via planning liberalisation has given volume housebuilders what’s called monopsonistic market power in local land markets. In other words, it’s created a buyer-dominated market. This has kept the cost of their land relatively flat while UK house prices continued to rise. 

[…] 

When market power in local land markets was combined with structural power over the state, we believe volume housebuilders were able to increase their profit margins rather than ramp up delivery to help the government meet its target of 300,000 new homes per year in England. Our research shows it was in the interests of the volume housebuilders not to rapidly increase their housing supply for two main reasons.

The resurgence of reactionary politics is entirely predictable and has been traced for a long time. Yet every victory or rise is analysed as new and unexpected rather than part of a longer, wider process in which we are all implicated.

The same goes for “populism”. All serious research on the matter points to the populist nature of these parties being secondary at best, compared to their far-right qualities. Yet, whether in the media or academia, populism is generally used carelessly as a key defining feature.

Using “populist” instead of more accurate but also stigmatising terms such as “far-right” or “racist” acts as a key legitimiser of far-right politics. It lends these parties and politicians a veneer of democratic support through the etymological link to the people and erases their deeply elitist nature – what my co-author Aaron Winter and I have termed “reactionary democracy”.

What this points to is that the processes of mainstreaming and normalisation of far-right politics have much to do with the mainstream itself, if not more than with the far right. Indeed, there can be no mainstreaming without the mainstream accepting such ideas in its fold.