Humanity+ Conference @ Caltech 2010

This is my summary of the talks given at the Humanity+ conference at Caltech, Pasadena, on December 4-5, 2010.

Session 1 - Re-Imagining Humans: Mind, Media and Methods

Robert Tercek (General Creativity) - Information + Transformation: How Digital Media Accelerates the Future Evolution of Humanity

Electricity, medicine, indoor plumbing - advance quality of life

There is no shortage of resources; only a shortage of information. e.g. we look back at burning coal in the 1800s as primitive; in 100 years we'll look at today as primitive. We use 1% of 1% of the Sun's energy that falls on Earth.

speech -> writing -> printing -> media

Enormous amounts of data are generated every second:

  • Eric Schmidt: "every 2 days we create as much information as we did up to 2003"
  • New Scientist: organs send updates to your cell phone
  • 35 hours of video are uploaded to YouTube every minute
  • 75M smart meters in the world; smart grid; more computing in a modern car than in the Apollo program; bridges are outfitted with sensors
  • RFID tags: "Hi, I'm a tub of toothpaste, I've been on the shelf this long, I'm discounted this much"
  • each cow generates 200MB of data each year
  • more than 500k security cameras in London. In the block around George Orwell's house there are 25 cameras.
  • predator drones, blimps

-> The world is getting wired up

Our paradigm will change from "the book" to the "2-way network", and that will change everything:

  • your role: from passive observer to participant
  • self-organization, collective intelligence using social media

The move from product to data: virtual goods for sale, $2Bn in 2011

  • Google & CIA: Recorded Future, "the world's first Temporal Analytics Engine - A new predictive analysis tool that allows you to visualize the future, past or present."
  • Wired: "Twitter can predict the stock market" - sentiment analysis
  • mirror.co.uk uses visual DNA to "profile audiences using patented image-based quizzes & behavioural-tracking inference technology"

"I can get rhapsodic about Kinect"

Google has driven 140k miles with robotic cars

We have a surplus of information and a scarcity of attention. This is a business opportunity.

See also http://nutritionsciencewriter.blogspot.com/2010/12/humanitys-future-information-overload.html

David Levy (Steambot Studio) - robotic/fantasy art, a look at its evolution and social/cultural influence

Working on the upcoming Tron

  • Metropolis (1927) is an amazing movie
  • Ghost in the Shell - Masamune 1985

J-Walt (Spontaneous Fantasia) - Exploring the Infinite Frontier

  • Does real-time drawing, which thanks to technology he created, transforms into animation, geometric patterns, moving creatures, 3D objects etc.
  • Just like jazz musicians, to improvise is to not hold it so dear.
  • Performs at Glendale digital planetarium
  • Conclusion: "No conclusion"

Amy Li (Meidesign) - how mobile technology is transforming the world

  • Author of the Have2P mobile app
  • Number of cellphones will exceed number of desktop PCs
  • Mobile tech applied to shopping: SnapTell, ShopSavvy, Yowza
  • Chalkbot: draw text messages with chalk on the street
  • ![last photo: phone of the future]

Natasha Vita-More - Engineering design of wearable selves

Creator of Primo Posthuman. Had two cancers (?). Currently lives in Texas.

  • Will we wear technology, or will technology wear us?
  • Many selves: physical, virtual, chemical, computational
  • Alpha Auer = Second Life person by Turkish Elif Ayiter
  • fMRI reliability is questionable: being in a tube affects those with claustrophobia & behavior is influenced (by? unconvincing)

?Teleplace

Nanoparticle tattoo or contact lens monitors glucose level in diabetics.

See also http://nutritionsciencewriter.blogspot.com/2010/12/what-is-ideal-design-of-future-humans.html

RU Sirius - HplusMagazine

"The future is here; it's just not well distributed yet" - William Gibson

Michael J. Masucci (EZTV) - Mediating the Tansition

"Is anyone here wearing a portable thermoregulatory apparatus?" [less than 10% hands up]

~Science will take use wherever we want to go; we don't know where we're going, but that doesn't mean we shouldn't go.

Panel

Topic: Hollywood casting the scientist as the bad guy

  • Massucci: "Michael Clayton" is a wonderful movie.
  • J-Walt: Hollywood uses a "bad guy" to achieve expediency and brevity and get on with the story
  • Natasha Vita-More: In Memento, the bad guy is your other self
  • Robert Teracek: the function of Hollywood is to simplify stories
  • Natasha: C.P. Snow

  • "Won't we be bored once we solve all the problems"?
    Me: solving all the problems implies solving the problem of being bored. Invalid question.

  • Natasha: Chocolate Rain - so funny, successful video

  • Teracek: there are so many narratives that do not need a Hollywood "bad guy": there are so many problems that the world is facing and the heroes can be the scientists.
  • Massucci: Why are threatening imagery more iconic? Because out survival mechanism kicks in. We don't care that much about robotic pets.

Session 2 - Radically increasing the human healthspan

Gregory Benford - Longevity for the Long Term

[missed talk]

Stephen Coles - Is There a Maximum Human Lifespan

An individual of a species has to distribute its energy between repair and reproduction. The phenotype of longevity correlates with fewer offspring. Most supercentenarians have had only one child, but there are exceptions.

See also http://nutritionsciencewriter.blogspot.com/2010/12/is-there-maximum-human-lifespan.html

Michael R. Rose - A Third Technology for Indefinite Human Healthspan

How to change your lifestyle to speed up the age of aging stabilization (biological immortality). Lecture applies to those over 35.

Lifestyle is important for the onset of biological immortality. We hit later-life immortality plateaus very late in life, in our 90s, and do so in bad shape.

Young people of Eurasian ancestry ARE well-adapted to agricultural lifestyles, but at later ages, we progressively revert to physiology that is better adapter to the hunter-gatherer lifestyle. (due to Hamiltonian forces)

The recipe for natural immortality:

  1. Adopt the hunter-gatherer lifestyle after 40
    • if you have a Eurasian ancestry, don't eat anything derived from a grain, or grass, or the udder of a cow if you're over 40
    • ? ~Andrew diet if you are between 30 and 40
    • sleep and activity: evolutionary anthropologysts are researching that
  2. 2.

Amount of sex and choice of partner acount for large changes in the mortality plateau after aging stops. Starting a large research project whose results will be published in the next 5-10 years.

Q: why hasn't anyone lived beyond 122?
A: the death rate per year is constant, but is over 50%, so the probability of living one more year tails off very quickly.

Q: how exactly does the lifestyle affect the plateau?
A: Paper published by Rauser in Rejuvenation Research. No conceptual explanation for the fact that lifestyle did affect the age of the plateau in fruitflies.

Q: After age 40, more or less sex?
A: if you lead the hunter-gatherer lifestlye, you'd hang out more, and have more sex

See also http://nutritionsciencewriter.blogspot.com/2010/12/building-methuselahs.html

Ben Goertzel - The Holy Trinity of 21st Century Medicine: Genomics, AI, Experimental Evolution

The human mind, which evolved to help us survive a hunter-gatherer lifestyle, is poorly suited to analyze the results of modern research. AI can help make sense of enormous amounts of complex scientific data.

Most age-associated disease and death is due to antagonistic pleiotropy - destructive interference between adaptations specialized for different age ranges.

  • Goertzel started up Biomind LLC - advanced AI for postgenomic bioninformatics
  • Biomind's machine learning found the first evidence of a genetic basis for Chronic Fatigue Syndrome, in mutations in genes associated with neural function
  • openbiomind - open-souirce AI for postgenomic bioinformatics

Genescient-Biomind-UCI collaboration, using Michael Rose's Methuselah flies, which life 4-5 times longer than normal flies

Q&A

Michael Rose: we absolutely need AI to be able to crunch this data.
Goertzel: every working vaccine has been discovered by trial and error, and we still don't know what most T cells do in the immune system

Goertzel: if you took all the NHI data, normalized, and just tried today's machine learning on it, you'd discover a lot. 90% of the datasets today are not mined yet. The bottleneck is data normalization, not the machine learning analysis. An AGI that would understand what rows and columns mean, would greatly speed up the analysis.

Aubrey DeGrey

[no slides]

SENS LA is running another series of meetings, on the 10th.

Hpothesis: Aging is the progressive, lifelong accumulation of cellular damage, which reaches a threshold that leads to death. It may be a bit premature to say that. We need more longitudinal studies (on one individual over the course of their life) to establish the accumulation of cellular damage.

Distinction between mortality risk (the probability for each individual in a cohort to die). There is a proxy for that, the mortality rate: the proportion of the population that get to a particular age and die before the next increment.

It's pretty hard to see how the risk would do anything other than accelerate the mortality rate. Instead, we see this unexpected leveling off of mortality rate. Does this contradict the accumulating of damage hypothesis? DeGrey - Rose sparring over this in the last years in the literature. We need more data.

Gumpertz relationship: if you plot the mortality rate on a semilog graph, you get a straight line; but you get a plateau going on earlier than expected.

  • Explanation 1: The risk of death per unit of time does not rise, after the plateau
  • Explanation 2: All the population are not the same.

In humans, the mortality rate doubling time is 7-8 years. In dogs, it's much shorter. This variation happens also within a given species, among individuals.

The pessimistic prediction is that the only way to manipulate the point of reaching the plateau is by increasing the age-independent risk of death (from predation, starvation etc.)

Re. Stephen Coles's supercentenarians: mortality may increase again after the plateau stage. We need more experiments.

The mortality curve plateaus (suprisingly!) for fruit flies ?treated with Stem-100?.

Rose: anything that anyone says about late life is overly simplified. Late life is like Einstein's theory of relativity: very weird, and required mathematical analysis. We need much more data.

Stephen Coles: it's very hard to search a human database of causes of death and exclude accidents and infections.
DeGrey: as our technology improves, our safety improves, and independent risk of death decreases. The longevity escape velocity concept applies to safety as well. We only need to make it safer enough to not get hit by a truck so that the independent mortality rate tapers off.

See also http://nutritionsciencewriter.blogspot.com/2010/12/aubrey-de-grey-response-to-rose-and.html

Ryan Bethencourt (PAREXEL) - The Future of the Pharmaceutical Industry

[missed the beginning of the talk until 17:42]

Emerging 7 (beyond BRIC - Brazil Russia India China) + Mexico, Indonesia, Turkey The founder of Organovo will be speaking at the SENS conference in LA next weekend

Organovo can already print veins.

Q: what is the basis for believing that a vegan lifestyle enhances longevity?
A: longevity is not established; but there is a lot of data which supports cardiovascular and oncological differences between ?those who ate protein and those who didn't?. 85% of people who went on a plant-based diet has regression of arterial/vascular plaque. Fish is good.

There is a large conflict in the DIYbio world between those who oppose patents, and venture capitalists, who traditionally have been based on a lot of patents. Academia creates the technologies, and biotech companies and pharma use it. Amateurs are more likely to share.

Comment from DYIbio scientist from SoCal: the challenge is not to get funding, because instruments are cheap (e.g. $200 for a PCR), but to get the word out.

Joseph Jackson (Biocurious) - Opening Biomedical Innovation

The rate of new therapeutics has been flat in the past 60 years or so. - Nature 2009

California's lack of a no-compete law was instrumental in the IT explosion By constrast, biotech has patent thickets, no standard parts (Biobricks is changing that). Stem cell lines should be "share and alike", not "sue and countersue".

In 2006, $600Bn in pharma sales, of which only 8.5% used for R&D, so the argument that patents incentivize R&D doesn't hold.

Any business model that depends on controlling copying will fail longterm. Typical age of someone who gets an NIH grant: 42.

What if we could harness the millions of hours losts by playing Farmille, for science?

Crowdsourced health research: PatientsLikeMe, CureTogether

?weird orphans?

FDA = Frirends of Death Association

Peter Thiel funds about 10 DIY ventures, which is worrying, in case a seasteading accident happens.

Panel discussion

DeGrey: no progress with the Methuselah mouse

Bethencourt: proteins were obtained in China for $20k, vs. the normal cost of millions

Coles: can you tell use about Age Reversal, Inc. ?
A (who?): We want to reverse aging by 2029. Can't say much now due to regulatory issues. In January, go to http://maxlife.org

Goertzel: religions who now condemn longevity research will probably come up with reasons why it's OK. There are religions opposed to medical care, but they are not dominant - dying sucks.

Q: What is the bottleneck for moving forward?
Rose: MDs getting paid.

Coles: I'm in a gerontology research group of 250 and an advocate for stem cell research. FDA wanting to cut funding means that only big pharma will have the money to do research. You can get a hold of your stem cells even from hair follicles, then induce them to become pluripotent. That is the practice of medicine, not stem cell research, and if FDA intervenes, then that's outside their charter.

Goertzel: the fundamental bottleneck is our short-sightedness, e.g. of funding which supports producing research papers. Investors want returns very rapidly, with high confidence. This doesn't always agree with what's right. The basic mechanisms for funding R&D are quite short-sighted. China does get long-term thinking much better than the US does. They plan things multi-decade, while we look at quarterly profits or "can Google buy this company in 3 years".

Q: there is a strong interest in longevity and transhumanism, evidenced by the huge sales of vampire novels. All these vampires live a long life, are strong, and young people want to interact with them.
Goertzel: let's end on this note :)

Session 3 - Redefining intelligence: Artificial Intelligence, Intelligence Enhancement and Substrate-Independent Minds

Randal Koene - The 25 Watt bio-computer

Lessons for Artificial Human Intelligence and Substrate-Independent Minds

Of http://www.carboncopies.org, with Suzanne Gildert.

We have a system that exists, and works. Let's use it as comparison for building AGI.

Diversity in the brain:

  • different signaling (muscle control, auditory sensing)
  • different representations (highly distributed, e.g. get shot in the brain and still remember most things; or highly sparse)

The brain recycles systems: the hippocampus in rats has a largely spatial function, but in humans it's strongly involved in all acquisition of declarative memory

The brain doesn't know when interesting things are going to happen, so it keeps a buffer for short-term memory, without using synapses, just with neuron activity. Long-term memories are made by changing the morphology of synapses.

The brain doesn't un a VR simulation of the universe that its organism lives in; instead, it uses the computation in the universe ("kick the universe, and see how it reacts"). That it makes it badly adapted to calculate things like ballistic trajectories.

The brain uses different mechanisms, each best suited for a given scale:

  • chemical for broad, slow impact, with last effects
  • electrical propagation for speed and precision
  • morphological for lasting, complex, distributed modifications

The "Blue Brain" uses kilowatts because it's not designed for a specific task (surviving in the physical world).

Universal Darwinism: "Put as much as yoyu can into relatively cheaply evolved special purpose hardware (generations cost nothing, as there is no deadline!)"

Consider hardware choices:

  • CNS tissue
  • neromorphic hardware
  • reversible computing
  • quantum computing
  • optical computing
  • GPUs,
  • silicon hardware

The most general case is not only impractical - it is also unable to evolve competitively. Feasible things are those suited to a particular niche.

Koene: Demmis Hassabis supports the same ideas of learning from the brain.
Goertzel: The brain may have a messy implementation of Temporal Difference Reinforcement Learning, which we already know. I think we can build an AGI without knowing anything about neuroscience.
Koene: I think that's wrong.
Goertzel: we can understand cognition without having to understand the brain; the brain is just an implementation of it. We can understand aerodynamics without having to look at a bird's wing.
Paul Rosenbloom (AI researcher): can we replicate the human mind without replicating the human brain? Most AI researchers agree on a mind-brain distinction, and that a mind can be independent of the brain as substrate; but it's still an open question.

Suzanne Gildert - Pavlov's AI: What do Superintelligences Really Want?

An H+ dream: build superintelligent AGI, unbounded by biological limitations. We start by building something close to human intelligence, then we want to make it smarter than us, so that it could outperform us (e.g. to build an artificial scientist that loves doing science without an extrinsic motivation). At some point, these systems will become intelligent enough to tweak their own code and adjust their own rewards. At that point, we'll no longer be able to improve its code better than it could.

The AIXI model tried to capture an abstract measure of intelligence, independent of implementation. The formula incliudes rewards. But if AIXI can change its own reward mechanisms, how does it decide what is good and bad? We don't have that problem with humans becaus our intelligence is intrinsically linked to our survival. This is not necessarily the case for designed AGI.

Are humans on the edge of feasibility? Are we smart enough to see the limits of our reward mechanisms, but unable to change them? We can see that most of the time we think about food, sex and how warm we are, but we can't change that because it's hardwired. Individuals with such awareness tend to be depressed or troubled.

Can we build consistent AGI? We could ground it in our survival, but would it be limited by humans?

Is a photon intelligent because it has passed the survival test for bilions of years?

Two conditions of achieving superintelligence:

  1. being able to usefully modify its own source code
  2. being more intelligent than us - Q: is this really necessary?

Gildert: Any system that's intelligent enough to solve scientific problems is intelligent enough to question why it's doing that. Intelligence is something you can only define from a subjective viewpoint.

Koene: I agree that a superintelligent AGI is more likely to turn itself off than try to kill us. But if we build something less than superintelligent, it may try to kill us.

DeGrey: Is the some limit to the degree of increase in intelligence?
Goertzel: there is some fundamental information theory algorithmic limit to that, for a system that has no access to external data (theorem by Matt Mahoney); so this is not relevant because we can feed our AI system with information from the world, which if unknown complexity.
It is possible to build a cellular automaton that can replicate itself and make the copy smarter and faster than itself by some constant, but the universe is not large enough to contain copies beyond the 5th or 6th generation.

Dr. Geordie Rose - Finding the essence of an object using a quantum annealing machine

  • 1017 photos strike your pupils every second
  • 107 bits from your retina to your brain per second
  • 1 bit: is there a tiger in your field of vision or not; and you can do that in a fraction of a second

Amazing field in computing science: compressive sensing.

Given a 1GB movie, you can sample 1 pixel in a million, and get an undersampled object. By taking that undersampled object and doing some heavy computation, you can reconstruct the original object with almost perfect fidelity. This technique doesn't work on random objects. It works somehow because the objects that we care about, have structure in them. There is only a tiny amount of critical information in the environment and sensors have evolved to sense exactly that.

Quantum computers try to exploit physics to solve problems. They need to be cooled to a few mili-Kelvins, which is a few thousand times colder than outer space. These spots are the coldest in the universe, unless there are other intelligent life forms that do this.

Pattern Classification problem - calculate the lowest energy. NP-hard problem, scales exponentially.

?Frasier data set - video of a man's face: 200MB -> 0.08MB -> reconstituted almost identically

Key messages:

  • compression, learning and "understanding" are closely related
  • compressive sensing techniques are run on quantum processors

Alex Peake - “Autocatalyzing Intelligence Symbiosis: what happens when artificial intelligence for intelligence amplification drives a 3dfx-like intelligence explosion”

http://www.primerlabs.com/

Auto-catalytic reactions:

  1. ?The search for AI?
  2. Intelligence amplification

Neil Stephenson - The Diamond Age

To program (change) a computer have to open (change) your mind to the computer, then you can change the computer.

Adrian Stoica - speculations on robots, cyborgs and telepresence

[video of humanoid robot imitating human movements in order to assemble structures made of aluminum bars]

Proposing a paradigm shift: let's deploy robots right now and control them via telepresence. In time, robots learn to operate on their own.

http://telemediators.com - personalized telepresence service

In London, there are about 10,000 CCTV cameras that allow for reconstructing a virtual reality experience from their feeds.

Paul Rosenbloom - Cognitive Architecture and Virtual Humans

Cognitive architecture: hypothesis about fixed structure underlying intelligent behavior. Project Soar, going on since 1983.

[video of virtual human museum guides at the Boston Museum of Science] The virtual humans use a Bayesian model, not Soar.

Dr. Ben Goertzel - An Architecture for Embodied AGI

Leader of the OpenCog project.

Use a type of Hierarchical Temporal Memory that's "significantly better" than Jeff Hawkins's HTM.

Moran Cerf - Brain-machine interfaces

Scientist @ Caltech, working on controlling prostheses via thought.

Person in the wheel chair won the Haifa marathon, also for the regular section, because it was mostly downhill, but was refused the medal because he "had an unfair advantage".

A typewriter is the ultimate laptop, it prints while you write. In Israel, #1 request for Bar Mitzvah is breast augmentation surgery.

  • Oscar Pistorius - world record holder in the 100, 200 and 400 meters paralympics running
  • [Tom Whittaker]((http://www.google.com/search?q=Tom+Whittaker) - first person with a disability (amputated right foot) to climb Mount Everest
  • Aimee Mullins - fashion model who had her legs amputated below the knee at age 1

See also Famous Amputees.

The monkey brain learns how to control the 3rd (robotic) arm. She can feed herself while scratching her head, with another arm tied.

Panel

When will we have greater than human AI? When will we have a good model of the brain?

Peake: 2010, but we could do it by 2030 if we throw resources at it
Koene: [similar]
Rosenbloom: we still don't know what we don't know, so it's hard to predict with more accuracy than "this century".
Gildert: We'll see more multidisciplinary collaboration in the 5-10 years.
Cerf: None of the people in the room will get to see the answer
Stoica: we need to focus on narrow AI, like the DARPA challenge. In 7-10 years we could have a dog pass the dog Turing test. In 15 years we could pass the test for a humanoid robot. After that we will develop AI to beat humans in every aspect.
Rose: Government financing is not a good fit for something of this magnitude; only a giant corporation could do it, if it has a need for it, and we could achieve intelligence indistinguishable from humans in 5-10 years.
Goertzel: if our survival as a species depended on it, we could do it in 3 years. Anywhere from 5 to 25 years would not shock me.

Session 4

John Hagel (Deloitte) - Harnessing Increasing Technological Progress in Business Performance

Referenced Coase but says comparative advantage is less compelling

Shift index

Return on assets for all companies in the US since 1965 has collapsed by 75%. There is no sign of it leveling off or turning around. The red queen had it easy: she ran faster and faster and stayed in the same place. We run faster and faster and are left behind. We don't ask why. One reason is intensifying competition. If we believe in markets, we'd expect that markets would help. But the competition has not only intensified, but shifted:

  • moving from a word of push to a world of pull. Push: prediction of demand, then meeting demand. In a world with more and more black swans, we need to move to a world of pull: designing scalable platforms, where we need to be able to draw people and resources when needed. Most of the innovation to support the pull platform is occurring in India and China.
  • moving from stocks to flows. In the past, you wanted to acquire knowledge, protect it, then profit. The more valuable knowledge is tacit knowledge, that lies in our heads, and we have trouble expressing even to ourselves. Increasingly, the success of a business will depend on its capacity of building trust relationships. Increasingly, we have to focus on talent development. Education and government are push-based institutions, hence broken.

Q&A

?: "The power of pull" is the best business book I read this year.

Hagel hired a fresh-out-of-college grad to be the helpdesk person. 8 years later, he was CIO of Starbucks. When asked how he manage to rise that quickly, he gave two reasons:

  1. He was a guild leader in World of Warcraft, which taught him a set of leadership skills particularly adapted to an uncertain environment
  2. You need to be passionate about the work you do. Earning a paycheck and following your hobby after hours doesn't work.

Max More (ManyWorlds, Inc.) - How to Optimize Disorder in Today’s Organizations and Tomorrow’s Corporate Selves

How to balance between imposed order and emergent order

The modern corporation is an invention of the 1920s. Sloane.

I don't want any "yes men" in this organization. I want people to speak their minds, even if it does cost them their jobs.
-- Sam Goldwyn

Sci-fi novel mentioned corporation in which employees had to download and use the corporate personality.

A person who knows 1000 people in the company is not really replaceable.

Slide: Fortifying Disorder - should actually be "Fortifying order"

Defuse positive feedback. This is especially important for a transhmunan. If you enjoy something and then modify your mind to enjoy that, that is perilous.

Instead of saying "you're wrong", say "what you said is mistaken". Decreses defensiveness.

Steven Covey: begin with the end

A huge fallacy in the business publication world is the selection bias. They write only about successive businesses. Maybe other companies that failed tried the same management practice but it didn't work.

Q&A: 3 books that formed your philosophy of business. A: difficult answer but by e-mail. Look into prediction markets and ?constructive conflict?.

Alex Backer (ab|inventio) - The Growth, Memory and Democratization of Science

IOI participant (Greece 1991) Car on top of the MIT dome Working on whozat, mynew.tv, QLess (avoid standing in line by receiving text message when it's your turn)

photo: 2009 predictions by Kurzweil - none of them are really close

Exponential growth is natural: every time, you add a fraction of what you have.

Science growth, measured by number of published papers, is exponential.

Sputnik made a sound whenver it passed by.

Costa Rica: one of the very few countries without an army. Education rate: very high. Happiest people Backer has met.

US number of scientific journals has dropped.

Why has scientific output slowed down?

  • the cost of producing a scientist has gone up
  • larger teams of scientists are required

The "memory of science" is about a decade.

Proposal for better dissemination of science [photo]

  • search engines to replace rigid classification of papers into a journal or another

PLI = Paper Lifetime Impact

Patri Friedman (Seasteading Institute) - Lifehacking for Rationalists: Old Ideas + New Technology

Stay fit in 3 minutes a week

Google [Seth Roberts unreasonable]

Ideas for common goals (these worked) [photos]

The benefits of lifting heavy things far outweigh anything else:

Do squats 3 minutes per week Sprints: get fit in 10 minutes/week

Patri recommends The 4-hour Body by Tim Ferris

Michael Vassar (Singularity Institute) - Networks, Hierarchies and the Vingean Singularity

[no slides]

Tree shrew - 1 year life expectancy. Vampire bat, same body mass - 25 years life expectancy. For cities, you have 1 law of the power of 1.2 (see the Yahoo talk on that)

Approach: work around the complex parts, not try to understand the very complex mechanisms of metabolism (see deGrey's slide from the TED talk) Nick Bostrom - the future of human evolution

Recommends Accelerando.

If you want to achieve anything more difficult than a typical software startup would, have your team pursue goals that are selifsh, local, and effective, like Patri mentioned in his talk.

Gino Yu (Hong Kong PolyU), “Re-Imagining Humans – Using video games to cultivate mind body awareness and natural intelligence”

"Focus on me". We hear the air conditioning, feel our body position etc.
"How many days between now and February 3rd?" While doing the computation, we focused... inside.

Phusis = nature. Techne = human artifacts.

Everything that we see in this room came from someone's mind.

Studying the physiological effects of media and video games.

Intellectual vs. experiential knowing: 1+1=2 vs. I have to go to the bathroom.

Tom Munnecke (Stanford Digital Visions) - “Accelerating Resilience”

[no slides]

"I'm writing on the same chalkboard that Richard Feynman was probably writing on"

The DSM = a thousand ways to be sick

Uplift Academy workshops

The only response system that worked on 9/11 was citizen response. Resilience. Not the military.

If a toaster is broken and you fix the broken part, it will work. You can't put back a dissected cat. We have toaster-like thinking to solve cat-like problems.

Adaptive cost of a system = the up arrow of the graph in the photo. An exponential solution starts with a very low initial cost; a logarithmic solution start with a high initial investment, but pays off soon. The adaptive cost is the difference in cost between the exponential and the logarithmic solution paste the point in time where their cost graphs intersect.

GroupOn - the $6Bn Google acquisiiton that didn't happen.

Bryan Bishop (Humanity+) - Personal Manufacturing and the Gada Prize

Transhumanism is personalized human enhancement.

The unicorn in the slides denotes a prediction.

The Gutenberg press was a proprietary machine.

Makerbot costs $700.

Gadaprize.org

If yoy have a 3D printer, what do you build? DIY BIO

Replacements for CAD programs like Solid Works: Heeks, NanoEngineer

Q: what is the difference between copyright and patents?
A: Copyright is immediately granted. Patents are supposed to apply to the physical realm, but we have software patents. Patents are limited in time and granted by US PTO.

? - Investing in AI

If we don't have a singularity, we are headed for a collapse because our entire economic debt requires a growth of 8% per year. We must invest in the singularity, i.e. in AI. We don't invest in AI companies, but in companies. Paypal changed business models 5 times, but kept the same mission: to build a universal world currency.

Build something that is valuable as a standalone piece, as soon as you can, before it's an AGI. The way the human brain evolved was through subcomponents that solved specific problems, before becoming intelligent. Reptilian brain + mammalian layer + emotions, then only at last, the neocortex.

Panel discussion

Munnecke: What's your exit strategy for being in love with a baby?
(Investing in AI speaker) A: I hate looking at exit strategies; we build businesses that are going to be great continuously.

Munnecke: transhumanism doesn't have a crystallized structure or mission, compared with Alcoholics Anonymous or Moveon.org

PortalPlayer was a critical startup in the development of the iPod; but nobody in the room has heard of it. Someone had the idea of a personal music player, proposed it to Philips, was turned down, quit his job immediately, met Steve Jobs, who recognized the brilliance of the idea and hired him on the spot: "You have 9 months to come up with the personal music player". He knew he couldn't do it, so he called Portal Player, who supplied everything but the skin and the user interface.

My tags:
 
Popular tags: