These are my raw notes from Day 1's morning session I think I have edited out any extraneous comments. h2. Jeff Jonas -- IBM p. Real time trip wires: black list + all other known possible problem makers and mix it into a mashup * Problem of enterprise amnesia: database silos, thief databases not merged with employee dbs or marketing dbs. * "Perception isolation" -- leading cause of enterprise unintelligence * "Perception integration" means integrated context * Want a system to observe and publish discoveries. * more intelligent systems created when you put data and persistent queries into the same data space. * notion of "perceptual analytics" * "relevance finds the consumer" * "when data and queries are treated the same then queries find queries" * observation that .6% difference between humans andchimpanzees, no wonder information systems are so screwed. * federated search hopeless for just in time context if trying to deliver enterpsie intelligence * must have enough features for context construction * can't bulk load data, must incrementally lean * "sequence neutrality" -- no matter the order of arrival of the data, the end state remains the same. * need to fix the past, fix connections as you load data. * notion that dreaming is deep recontextualisation * has a blog * how to handle ambiguity -- add more observations h2. Bill Janeway, Peter Bloom, Tim O'Reilly * bloom: that there's three themes: ** extracting signal from noise ** reduction of latency * That NYSE trades average 30 ms, hedge funds demand 1000 trades/second * Even in the era of the internet location matters, example of company moving transaction server closer to NYSE to shave 10ms off tp. * Example of India prohibiting electronic communication between markets. Solution was to have guys watching two screens and typing in to two systems at the same time to do the arbitrage. Ie, Mechanical Turk * generation focused on transactional efficiency * war game over exposure to quais-proprietary information, who owns that and what to do with that? * notion that hedge funds give up their "clickstream" to prime brokers, hedge funds use alternative trading systems to try to hide their trading data from others * preserving anonymity by breaking apart transactions across multiple markets * "dark pools of liquidity" * "tradestation" -- all of the former proprietary data is available now in consumer package for $100. users can trade models with each other. darwinian evolution of trading models. * reminder that in answer o the 87 crash due to program trading imposition of hard stops. * slow down of markets during crises, is it a problem or a way of enforcing liquidity? * removing latency means there's less time for you to reconsider your decision if it was a mistake h2. werner vogels, cto of amazon * "web scale computing" , building your business around ideas and not resources * sales pitch for amazon's services * computing, storage, queuing * pay-go * GWA for real systems * claims 99.9% availability * amazon machine image -- how does this compare to media temple's grid service? * sharing of machine images by community * switch to rightscale.com demo h2. Jane McGonigal p. url: www.avantgame.com/happiness . hacking the real world to be more like a game p. What are the design strategies for showing users alternate realities? p. technologists as happiness hackers h3. future forecast * quality of life as the primary metric for evaluating everyday technologies * positive psychology as a principla explcit influence on design * public expectation that tech companies have a vision for life worth living h3. science of happiness p. "three realms of happiness" * pleasure - satisfaction * engagement - immersion, responsive systems * meaning: having a powerful role in the world at large * how does engagement in second life, games, influence real life? * ubiquitous, alternate reality games, activating the world around us * example: ministry of reshelving (flickrtag:reshelving) move "1984" from fiction to U.S. history, current events, etc. * example: assassin with random acts of kindness. * using games to teach interaction in the social environment * "world without oil" alternate reality for the public good. 4/30/07 * worldwithoutoil.org h3. "the new games are supergames" * super sized, massive scale, lot of people engaged, online collective intelligences * superimposed (hybrid experience) * suerpheroic (real place, not role play) * supercomputing (parallel problem solving and collective action) * notion of being part of something bigger * one common design trait is that there is a call to action to work towards a larger goal. * search for shared meaning * ubiquitous games more popular with female players than you see in video games h3. happiness measures * UPenn measures, via questions/statements: ** "I am optimistic about the future ** I feel close to most people even if I do not know them well ** I feel that ....this makes the world a better place. h3. Six happiness classes * wisdom * courage * humanity * justice * temperence * transcendance h3. do our technologies pass the "death bed test" * happiness gained by assessing life backwards * How will the time spent with our technologies look in retrsospect? What kind of life will our technologies have enabled, supported? bq. Invest a portion of your timeenergy and resources towards undersand and innovating happieness, it's the new capital. p. make your technology not only feel good but also do good and expose good ...two more that i didn't catch, but slides at http://avantgame.com/happiness h2. jeff hawkins * cofounder of numenta * _why can't a computer be more like a brain?_ * have no computers today that you can show a picture and say what is this, where humans can do * no computers that can understand language or talk like a human h3. why are computers unable to do what brains do? p. suppositons: # computers are not powerful enough # brains are too complex to understand # brains work on quantum principles # brains are "magic" * representations built hierarchically, need to build model of the world * built from the sensory data, learned. h3. hierarchical temporal memory (HTM) * a theory of how neocortex works * HTM also a technology ** hierarchy of memory nodes ** learns by exposure to sensory patterns h3. how does htm work? # creats a model of its world # recognizes new patterns # predicts # generates behavior, * numenta platform for intelligent computing * kind of losing the thread here. * demo of pattern recognition. p. applications against precise timing or high order temporal data. See http://www.numenta.com
Posted in ETech
Copyright 2002–2011 Artific Consulting LLC.
Unless otherwise noted, content is licensed for reuse under the Creative Commons Attribution-ShareAlike 3.0 License.
Please read and understand the license before repurposing content from this site.