Exponential View

Azeem Azhar’s Weekly Wondermissive: Future, Tech & Society 

We’re at the mercy of your inbox filters, and they’ve been quite harsh following our transition from Mailchimp to Substack, a newsletter delivery platform. If you find this issue in your spam, please do the following:

1. drag the email from your spam to inbox,

2. add my email to your contacts,

3. mark the email as important or star it.

Thanks!

Dept of podcasts 🎧

The Exponential View podcast is supported by Spotify.

Hundreds of years ago you had a community of monks who had been stewarding civilization through the dark ages and had been keeping knowledge alive. They had access to privileged information, privileged knowledge, and privileged institutions. Then the printing press came along, and messed everything up for them. I think that the AI research community is going through this period now, where they're figuring out which of them are monks, which of them are peasants, which of them should be doing printing presses, and which of them should be keeping information private.

In the last episode of the Exponential View podcast in 2018, I discuss the state of artificial intelligence, the geopolitics of technology, and Edison’s elephant with Jack Clark, the Policy Director at OpenAI.

Subscribe to the Exponential View podcast and get the latest episode:
Spotify (recommended) | iTunes | Soundcloud | Stitcher | Overcast | Breaker

Reviewing my 2018 predictions 🔎 

At the end of last year, I published some predictions which were pulled together with the help of many readers. How did we fare?

Broadly speaking, we did very well. The strain on the international political economy, caused by the Second Great Divergence, as technological capabilities accelerate away from existing social systems, exacerbated. Facebook’s altercations with many governments being an exemplar. But we’ve also seen India’s government take steps against global e-commerce players. I predicted that Silicon Valley would hire outside fixers to solve these problems, or at least, improve their optics. Britain’s Europhile politician, Nick Clegg, starts a new job at Facebook in the new year. And we’re increasingly seeing the spotlight shine on Amazon and its control of commerce on its platform.

We also, roughly speaking, got the trend of innovation and scaling outside of Silicon Valley right. Chinese companies continued to scale very fast. Some like, Toutiao, delivered novel experiences that became the envy of the best in the West. Deep tech firms effloresced in areas like post-von Neumann compute, computational chemistry and other fields. Many outside of the Valley.

Our AI themes: more money, increased corporate prioritisation, focus towards human augmentation, progress in fair and transparent AI, and others were well represented this year. AI is booming, as I discuss with Jack Clark in the latest podcast (above). Equally, our calls on autonomous vehicles, cyber attacks, augmented reality, venture investment in health-tech and ad-tech fracas played out well.

What I got wrong was fascinating.

I thought that crypto technologies would start to show their utility this year. But it seems that true utility is still a few quarters away. I was ambivalent about whether the speculative bubble would pop, and thought money would probably continue to flow into crypto assets as institutions prepared to introduce new crypto products. In fact the speculative bubble popped, and fraud almost rivalled greed in its prevalence. The result was a retrenchment by traditional institutions from these spicy products.

The most well-funded crypto firms (even in the enterprise space) started to retrench and cut costs. The absence of evidence of real use cases was replaced by an evidence of absence of real use cases. Next year will likely be more productive for blockchain, as the dust clears.

I also underestimated the rottenness of Facebook. I’ve long drawn attention to the problems created by Facebook’s monoculture, its denial of an editorial role, its lackadaisical attitude towards the pollution caused by its attention model, or its role in debasing the public space. But I simply didn’t understand quite how deep this would run, and how it would spread across so many aspects of its business.

Facebook’s turpitude is legion: occluding honest debate and investigation via legions of mealy-mouthed communications; regularly faking its metrics (for its advertisers); its imperial powers of censorship and control, swung by the political axe of the public policy team; its comical belief that it made real progress in addressing its structural failings.

The firm continued to disgrace itself, having almost fully reversed its Midas Touch, in ways I didn't predict.

I also wrongly predicted increased Russian disinformation during the 2018 US mid-terms. Some argue that the Russians largely sat out those 2018 elections. What we did see was more evidence of disinformation and computational propaganda in many other geographies, promulgated by a wide host of actors.

In short, I think we, with the help of many readers, did reasonably well in reckoning on 2018.

I’ll endeavour to put some predictions together for 2019 for next week’s issue.

In the meanwhile, the rather brilliant Isaac Asimov put forward some predictions for 2019 back in 1984. They are worth reading.

Top 10 most popular stories of 2018 💯

10: Michael Osborne’s presentation on automation at work.

9: Michael Liebreich on pro-growth environmentalism.

8: How Microsoft is succeeding despite the end of Windows.

7: 2018 as predicted by people in 1918.

6: How data science and interactive notebooks may represent the future of the scientific paper.

5: How to predict a technology’s commercial success.

4: Human history in one chart.

3: Silicon Valley’s soul-sucking machine.

2: The Apple Watch is a bridge to the future.

1: Ark Invest’s Big Ideas for 2018.

End note

It felt like 2018 was a really dark year.

A dark year for politics, as populists continued to win elections around the world with divisive manifestos built on walls and scapegoats rather than bridges and inclusion. For those of us in Britain, the fact-light & insalubrious Brexit debate is an object lesson.

It was a dark year for the technology industry, whose behemoths are still trying to get to grip with their new systemic role in the world. Microsoft and Apple, two of the oldest, are doing a better job than the whippersnappers, like Amazon and Facebook. Big Tech feels more like a Gordon Gecko cigar, than Steve Jobs’ bicycle for the mind.

But equally, with only two business cycles left to get us to a net zero carbon emissions world, I find it profoundly depressing to see so much time, attention and talent being poured into firms like Epic Games, which makes Fornite, and Juul, an epidemically successful peddler of nicotine. Our future is not going to be assured by doing the Floss while huffing a Juul vape.

As I have learnt a little about early-stage investing with my friends at Kindred Capital, I have delighted in seeing more entrepreneurs wanting to tackle more meaningful problems: like female health, non-meat proteins, improved software tooling for AI, or tackling sustainability questions. And the amazing progress we've seen in the broad field of AI from research shops like OpenAI and DeepMind gives us cause for optimism. More so is the speed with which the insights of expensively-discovered avenues of exploration can move into more widely-available open source techniques, for all to use (Read this wonderful New Yorker essay on AlphaGo and LeelaZero.)

And equally, I’m starting to see some interesting intellectual commonalities emerge, especially around the tricky question of how we create economies which can halt the carbon-crisis while ensuring well-being. In my discussions with Kate Raworth (a sustainability economist who is roughly-speaking left-leaning) and Michael Liebreich (a climate change expert on the right), I saw more in common between them than I saw apart. A slew of economists and policy people, such as Bill Janeway, Mariana Mazzucato, Diane Coyle, Matthew Taylor and others, are reinvigorating the importance of the state in guiding innovation for social rather than individual purposes. (Many of these conversations are available on the podcast series. If you haven't tried podcasts, I recommend you do.)

And so, there remain reasons to be, in Paul Romer's words, "conditionally optimistic". (This year had many isolated bright spots as these fifteen charts show.)

Conditional optimism is not believing that if the market signals it or the customer demands it, then it must be right. Conditional optimism is not the bare naked solutionism of a lean startup powered by venture capital and a cheering demo day.

Conditional optimism is not "not complacent optimism. Instead of suggesting that we can relax because policy choices don’t matter, it suggests to the contrary that policy choices are even more important than traditional theory suggests."

2019 will be another challenging, difficult year. It is always darkest before dawn. But dawn will, ultimately, break.

Happy New Year!

Azeem

P.S. Saying Happy New Year via this tweet or LinkedIn post would be most appreciated! 🥂