|May 24, 2015||Public post|
IBM Watson a leading candidate for a general AI; stealth robotaxis; synthetic meat price collapse; disrupting lawyers & insurers; Dr Moravec; robot sex, bubbles and Blackberries
Still in beta: But please do forward me the names of two colleagues you think would enjoy reading this.
Dept of the not-so-far future
🔮It has been 4 years since IBM’s Watson won at Jeopardy. He is growing up. Should we be scared? IBM is pushing Watson very hard across hundred of different business domains and it is solving problems in text analysis, image processing, healthcare and finance already.
🚕Meet Zoox, the stealth robo-taxi firm taking on Uber and Google. (Backed by Steve Jurvetson of DFJ)
Dept of future food
Feeding nearly 8bn humans comes at huge cost. It will get harder and harder as the years of antibiotics, genetic modification over-fishing and over-farming take their toll. Rethinking the food supply chain is one area where gains are much needed.
Remember the lab-grown burger that cost around £200k per patty? In a year, the cost has collapsed 30,000 fold 💥 putting it around £50 per kg or close enough to economically viable.
It isn’t just artificial meat that could upend the food supply chain. New techniques like 🏦vertical hydroponics are increasing the efficiency of farms, and allow the sustainable development of large-scale vertical farms within urban centres. For example, artificial sunlight from precise LED-lighting, makes them 100 times more efficient per square foot than traditional farms..
Dept of next-to-be disrupted
Three industries have resisted disruptive new entrants include hedge funds, insurance and professional services. No more.
🗿Yann Ranchere of Anthemis Group pens a fantastic essay outlining the technology innovations upending the $1 trn insurance industry. Long, thoughtful must read.
Rachel Botsman looks at how platforms and marketplaces might kill 💀 the professional services firm. Lawyers and management consultants have resisted platform based disruption, boutiques like Eden McCallum, or lower-tier platforms like UpWork not-withstanding. New startups take an Uber-approach of vetting and rating talent and making it accessible via a platform.
The ‘alpha’ that quant-strong hedges funds once had may be under threat for super-commoditised hardware (as discussed here) and also advances in automated learning.
Dept of Dr Moravec
Moravec’s Paradox, that what is quotidien for us is hard for a computer but what is hard for us is easy for a program, has characterised AI & robotic progress for the past 40 years. Starting to change.
A Berkeley research team has applied a deep-learning technique to train a robot to use trial and error to open bottles and build Lego. This view on Kurzweil AI is a slightly more technical description.
In fact, an AI has learned to spot depression about as accurately as a trained professional. (But obviously much, much more cost effective than a human psych.)
Another piece of the puzzle is to provide systems that allow the pursuit of hierarchical goals. Many real-world problems solved by humans include delayed roles. Here is nice post by a Czech AI gaming company developing a delayed reward learning algorithm.
Short morsels for dinner parties
🎈 Don’t mention the B word. Silicon valley doesn’t do bubbles.
Must read 💫. An easy explanation of the top 10 data mining algorithms. Allow yourself to look smart at Board meetings.
💋💘Sex and love in the age of robots. It is bound to happen; and when it does it’ll reshape our relationships with robots (and potentially with sex.)
An 👶 AI gets smarter than a toddler. (For image recognition - but it’s clearly not the same AI that Flicker used.)
Flickr 🙈 started to auto-tag photos using an automated classifier. Hilarity and offense ensues.
The calamitous fall of 📱Blackberry. They had the world and threw it all away in a textbook case of being disrupted.
What to learn to stay ahead of computers. Robert Schiller argues students need to be taught ‘the art of living in the real world.’
A hedge fund is using big data analysis to challenge poorly awarded pharma patents. With real results.
On the need for a minimum basic income. Very American-centric but not a bad argument in favour of this policy change.
Bitcoin’s decentralised decision-making system needs to collectively act to raise the block size (and allow the network to scale.) Interesting governance challenge.
This is issue ten of my digest. Really pleased to have you with me.
Please forward this on to friends, or send me names of people you think would enjoy this. Email is firstname.lastname@example.org