Podcasts: starting points for learning about AI

Stuart Russell on After On with Rob Reid

A thought provoking interview and excellent introduction that sounds a note of caution about AI and gives good reasons for doing so.

Highlights include:

  • some interesting stuff about how people and robots process goals and the huge number of actions and priorities that make up a single ‘simple’ action (around 32 minutes);
  • Discussion about recent progress with AI learning to play real-time strategy video games that are far more complex than chess;
  • A definition of ‘beneficial’ AI and some other nuances beyond standard ‘general artificial intelligence’ around 43 minutes;
  • A brilliant illustration about robots cooking cats at 1 hour and 16 minutes.

Rodney Brooks on Econtalk

Brooks is less concerned, and takes an ‘AI will take a lot longer to develop than anyone thinks’ approach to the topic, with some good points about how developing AI forces us to clarify our own ethics and priorities.

Start the Week with Yuval Noah Harari

Harari paints an unsettling picture of a post-human future.

Amy Web on Econtalk – Artificial Intelligence, Humanity, and the Big Nine

On my hit list. I’m a Russ Roberts fan and expect this will be a useful addition, in particular on “the implications and possible futures of a world where artificial intelligence is increasingly part of our lives.”

See Also

Resources in WtF from Kevin Kelly, Tim O’Reilly and James Gleick,

Bootstrapping the non-profit organisation Rule 7 (part 4)

This is the seventh-and-a-quarter post in a series applying Seth Godin’s rules of bootstrapping (see also here) to building a non-profit organisation.

Rule 7: Charge a lot (but be worth more than you charge)

A last argument for applying this rule in non-profit context is that if your clients pay, your resources go further, and you can serve a lot more people. We covered this principle in Rule 3 (see ‘Eager to pay and scaling the non-profit), but it’s worth repeating here.

I’ve just been listening to an interview with One Acre Fund‘s Andrew Youn on Rob Reid’s After On podcast where Andrew spoke about the importance of a revenue model in their work, where the farmers they serve receive credit, but ultimately pay for the services they receive:

Rob Reid: Some people might say that these folks are extremely poor, why don’t you deliver these services for free? Part of it is that there’s only so much money in your organisation, and that 98.5% payback means that you’ve got a lot more dollars put to work. What percentage of One Acre Fund’s annual budget comes back to it through repayments?

Andrew Youn: Most of One Acre Fund is core program delivering all these services… Within that core program, about 70% of our budget is covered by farmer payments, and 30% from donors.

RR: You’re literally serving three times as more as many people as you could if you were a purely charitable organisation.

AY: It makes us so much more cost effective… we can serve three, four times as many people by charging for our services. I think it also makes us a little more beholden, as an organisation, to the customer that we serve. So we use, for example, repayment as a customer service quality metric. [see Rule 1 for more on this idea]

After On Podcast, Episode 35

.

So here’s the final reformulation of Rule 7:

Rule 7 of bootstrapping the non-profit organisation: Find the right price (and be worth more than you charge)

Hybrids (2): combinations and connections

New ideas and technologies are often hybrids. Sometimes we take quantum leaps and invent entirely new technologies, but more often they seem to emerge at the intersection of existing ideas, tools or ways of doing things.

Walter Isaacson’s The Innovators: How a Group of Inventors, Hackers, Geniuses and Geeks Created the Digital Revolution is full of examples of how this happened by accident and design throughout the evolution of modern computers.

Tim O’Reilly has called this ‘combinatorial innovation’. Here’s Rob Reid describing it in an interview with O’Reilly:

Combinatorial innovation is taking completely disparate technologies that have arisen, and weaving them together in ways that create tremendous and unanticipated new things which really magnify what society’s gaining from those new technologies.

So you might say with Uber and Lyft, we’ve suddenly all got GPS in our pockets, for reasons that have nothing to do with ride handling, and we’ve got this mobile payment system that Braintree or Stripe or whoever created for completely unrelated reasons, and suddenly they combine into something society shifting that no-one saw coming – ride sharing, in this case.

You almost get amazing things for free that used to be impossible or wildly expensive by coupling together a few other new things that just kind of happened to be lying around, so to speak. And maybe these new things can become an ingredient to something even more amazing, which may create a lot more jobs and social good.

I think the example you use with Uber and Lyft, is that even in a worse case scenario from a jobs standpoint with self-driving cars… for example… a lot of jobs will be displaced… but the cost of a ride might also come down by 80%. So while there’s a lot of economic dislocation , a whole realm of new services can arise that are based on the sudden, extraordinary affordability and ubiquity of transportation, much as the abundance and cheapness of wool [in the industrial revolution] led to the rise of fashion. The world laid off a lot of weavers, but through combinatorial innovation a whole slew of opportunities arose.

Rob Reid, talking about the ideas of Tim O’Reilly on his After Hours podcase

Just afterwards, O’Reilly said this.