How does Apple do AI? Does big data mean the end of free will? How will electric vehicles rewire our economies? What makes that elephant chart is so terrifying? Who is really responsible for carbon emissions? How do chimps co-operate? How many times can you fold a piece of paper? What makes Col Sanders’ secret recipe so finger lickin’ good? So many questions to stimulate so many conversations. Enjoy!

🏖 Back from your summer hols? Please recommend us via the Twitter  | the Facebook 

Dept of the near future

😮 Big data and the end of free will. STUNNING essay from Yuval Hariri. “As long as you have greater insight and self-knowledge than the algorithms, your choices will still be superior and you will keep at least some authority in your hands.” h/t @shehnaz_s

🚘 The massive impact of electric vehicles on our economy. MUST READ It’s going to be sudden phase-change not a gradual one argues Michael Liebreich. EV sales are growing faster than expected and the change will affect all parts of the modern economy.

🍏 How Apple uses deep learning. Steven Levy gets a GRIPPING exposé on the Californian tech company, unveiling some juicy new details. (See more discussion below.)

🐱 The ’concept of cat face’. Paul Taylor on machine learning in London Review of Books. Beautifully written and ACCESSIBLE.

**💥 “**Technology has gotten so cheap that it is now more economically viable to buy robots than it is to pay people $5 a day.” EV reader, Kaila Colbin on why the elephant chart is so terrifying. GREAT READ

Dept of artificial intelligence (Apple special)

(_Bit nerdy - still great stuff below this, so scroll past if you have to.) _The depth of Apple’s commitment to AI has started to manifest itself since the summer. Stephen Levy’s essay manifests those efforts in greater detail than before.

We touched on some of these topics in Exponential View #67 back in June when I wrote: “Apple is going to continue its investments in improving UX through technologies in the AI stack… Expect more AI in Apple’s products. But I would be surprised to see large-scale open source efforts, of the kind we have seen from Google or Facebook. Open source has rarely been Apple’s bag.”

What is new in Levy’s piece is more granular details which include the following nuggets.

• Apple runs a neural net locally on the iPhone.

• This neural net weighs 200Mb and trains itself in real-time, but especially overnight, using the GPUs in Apple’s iPhone device.

• Apple cites owning the silicon design (from the far-sighted acquisition of PA Semiconductor, I guess) as a driver of improved learning performance.

• They replaced oldskool voice recognition  (hidden markov models) by a deep learning approach back in 2014.

• Apple uses third-party sourced data to generalise training of things like photo recognition. This happens on-device.

Given the level of Apple’s acquisitions in this area (see 💡 my tweetstorm from May 2016 which covers this, and was prior to the Turi deal), their recent senior hirings and more than 280 job openings for hardware & software engineers with machine learning experience it is reasonable to say Apple is going for it.  (Fascinated to learn more about “Proactive Intelligence”, their new AI-enriched interface paradigm which sounds a bit Weave.Ai/Google Now like.)

One open question raised by the Levy piece is whether Apple’s mental model around privacy is a bug or feature when it comes to artificial intelligence. Apple doesn’t share user data. And Apple’s global models are built not on this shared user data but on externally & expensively sourced data. And Apple doesn’t seem to send much user data back to the cloud to be learned from on super deep networks running on GPU clusters.

The traditional argument would be that it is a bug. Leveraging data network effects allows you to build better, more defensible products faster. Tesla’s network learning (EV#31) is a great example of this. As is Facebook’s capability in face & object detection. And keeping things on a local GPU denies your neural nets of the value of lots of GPUs (particularly for training).

The counter argument would be that user-privacy may increasingly be a differentiating feature which allows you to sell more stuff. Apple is wealthy and paying for tons of training data doesn’t make a dent in its cash pool. And, in any case, model performance often tends to a limit beyond which additional training data doesn’t help you.

Here’s my fast take on this. Apple’s approach to user privacy is may start to look more like a bug than a feature but it may not make a difference right now.

• Externally sourced training data can’t keep up with novel use cases generated by real users. So your external training takes a long time to improve your overall performance. An Apple car training locally will generally have worse training data than a network Tesla whose models draw on edge cases from across the world. Worse performance means a worse product means worse market share means…

• Their introduction of differential privacy which Levy discusses and we alluded to in EV67 suggests they see the value of data network effects and are finding a way to grab that data while staying true to the user privacy promise. What I don’t know is whether differential privacy provides sufficiently good data. I’d recommend reading this essay at High Scalability which looks in more depth at deep learning in Apple Photos and differential privacy.

• Andrew Ng, Baidu’s deep learning czar, has pointed out that deep learning performance doesn’t seem to flatten out as you add more data. (EXCELLENT SHORT PIECE) You can just make the network deeper and the model continues to get more performant.

• Consumers won’t care. For better or worse they won’t care enough especially when given the choice of products that feel more ‘magical’.

Right now (and perhaps for the next few years) this probably won’t hinder Apple. But their approach to user privacy might start to hurt the user experience they can deliver. Now that would be an interesting tension.

Elsewhere

• 😞 AI’s research rut. Olga Russakovsky on the problem of homogeneity amongst the stakhanovites of the machine learning industry.

• AI’s beat doctors at cancer prediction.

• Facebook open-sources DeepMask, its better than human object segmentation network.

• Facebook’s “Trending Topics” team, which was accused of political bias earlier in the year, has been replaced by automated headline writing code. 

Dept of climate change and renewables

🔥🔥🔥 The climate news really doesn’t get much better but I guess acknowledgement is the first step to action. Just 90 companies are responsible for the vast bulk of carbon emissions, argues geographer Richard Heedie. **STUNNING VISUALISATIONS. **Mostly oil majors and many, like Exxon, knew about the risks for close to half a century. (See EV#28.) Should those companies start to pay damages?

It is going to get vastly hotter says this essay with pretty, but unappealing graphs.

• The third pole, the huge ice asset high in the Tibetan plateau is melting fast. It will directly impact water supplies for 160m people and affect water and climate for nearly 2bn more.

• The North West Passage is ice free for the first time in a long, long time.

• The Arctic will be ice free next summer.

• New melts in Antarctica mean permanent loss of ice and arrival of thousands of new lakes, which further exacerbates ice loss.

Exponential View Bay Area Meetups

🌉 I’ll be in the Bay Area in November. If you would be interested in joining me for a one-to-one, dinner or high tea, then please fill out this form.  I’ll confirm my dates closer to time.

I am particularly interested in connecting with firms, founders or product leaders working in marketplaces or two-sided platforms; or anyone involved in AI, future of transport, virtual reality or blockchain-based services. Happy to meet with VCs or academics in those sectors too.

Let me know! Be great to connect.

Short morsels to appear smart at dinner parties

💡 What is New Economic Thinking? Amina Silim on how evolutionary, complexity and behavioural economics are coming together to set the neoclassical models to rest. Good primer.

👽 The Nature paper on the habitable zone planet, Proxima b. (What my Twitter followers think.)

🐒 Like humans, chimps reward cooperation and publish free-loading.

💸 The clear point when women’s wages fall behind men’s. (After the first child.)

Brazil’s globo.com streamed 5.5 Terabytes of live video to Brazilians during the Olympics.

Voice recognition is now as good as typing, says Baidu. (I use Siri for dictation quite a lot.)

Details on the Israeli crypto firm, NSO, and their zero-day exploit of the iPhone. (And the hack of a Bahraini human rights activist.)

🐣 The secret recipe for Kentucky Fried Chicken seems to have been leaked.

∫ How many times can you fold a piece of paper in half? It depends.

End note

I often read Yuval Hariri’s work, having been extremely impressed with Sapiens. Regular readers will know I refer to his essays frequently.

If anyone knows him well enough to do a warm introduction, I would LOVE to connect with him. Thanks!