Webtalk: Realistic Routes to Substrate-Independent Minds by Wikipedia:Randal A. Koene, carboncopies project/teleXLR8 the 17th of July 2010
Recordings, notes timestamps were in CET, presentation slides should become available.
- 19h00 ...irony
- interested in infrastructure/substrate reliability and security yet having technical problem joining the conference
- note for Programming#SoftwareStack that it was a stack trace that allowed to pinpoint and solved the VM problem in (nearly) due time
- 19h38 (finally joined)
- input (non-invasive, invasive, ...)
- Ted Berger cognitive prostheses
- 19h45 on brain limitation and constraints
- short and long term-memory
- cf my AFT prepared talk
- memory limitations imply complex (highly multidimensional) correlations limitation
- 19h50 environment risks
- pace of evolution
- repair and defense mechanism
- 19h55 whole brain
- first order, second order quantification
- 20h00 minimum data that retains information theoric mind
- ~minimally sufficient neural correlate?
- 20h04 function and structure
- 20h10 transitions
- how to get there
- technology becoming part of GPS
- lifelog, is it sufficient?
- in-vivo through micro-electrode arrays (MEA)
- 20h20 beyond the cutting edge
- 20h23 regional significance limit
- 20h27 summary
- experiment, demonstration, analysis
- satisfy ASIM objectives
- 20h32 Q&A
- organization and upcoming talk (and its potential broadcast)
- comment on Kurzweil computational emulation requirements
- realistic time line?
- depends on the resources injected
- typing points?
- proof-of-concept (even at a small scale)
- Ted Berger (but currently not up to date)
- what is ASIM?
- needs-based network
- Penrose and QM (later on question on microtubules/Penrose–Hameroff ‘Orch OR’)
- organic vs inorganic substrate
- both can carry computation
- each can be better based on the environment targeted
- several remarks on vocabulary used
- problems of "bad PR" thus recommending terms like "whole brain emulation"
- comments on Yale Study Shows Electrical Fields Influence Brain Activity July 14, 2010
- optogenetics and consciousness
- in particular Ed Boyden work consciousness
Personal motivation for attending
- understanding the trend better
- discovering how I can leverage it
- how do my projects fit in
- sharing knowledge based on my specific background that might not have been integrated so far
- live migration
- network topology and latency problems/opportunities
- energy requirements and against "substrate idealism"
- my mind has been uploaded, now what? (the case for "infrastructure as code" paradigm)
- can that datacenter crash?
- do I rely on the datacenter administration? energy supply?
- putting aside the physics and hardware problems, how does the law protect me? how does software license protect me?
- can the software locked-in problem be extended to this topic? what would be the consequences?
- can I re-locate myself?
- at what risk? at what cost? at what speed?
- how do I leverage that new substrate?
- what are the limits that I didn't have before?
- security? DDoS? botnet? (security is also "Substrate-Independent")
- problem of ransom?
- if those questions are ignored as "technical details" what are the chances of the first organisms to actually survive?
- is the learning curve worth early failure?
- even if it is yourself?
Groups and activities
- Own projects
- Key topics
- Reading notes
- Talks watched earlier WithoutNotes
- The China-Brain Project: An Evolved Neural Net Module Approach, AGI-08, May 2008
- Transhumanism: Enabling and Transcending the Human Brain by Susan Schneider, Penn Media Seminar on Neuroscience and Society, University of Pennsylvania, April 2009
- Building Brains: The Molecular Logic of Neural Circuits, ResearchChannel, 2008
- How the Brain Invents the Mind by Rebecca Saxe, MIT World June 2009
- Whole Brain Emulation: The Logical Endpoint of Neuroinformatics? by Anders Sandberg, Google Tech Talk May 2010
- Building a Circuit-Diagram for the Brain by Jennifer Raymond, Stanford University 2009
- Learning and Memory: How it Works and When it Fails by Frank Longo, Stanford Mini Med School March 2010
- Seeing Is Believing by Carla Shatz and Mark Blumenkranz, Stanford Mini Med School February 2010
- ~1h45min neuron and silicon, artificial synapse chip
- Bionic and Bio-ionic Neural Interfaces by Luke Theogarajan, UWTV 2009
- Semiconductor-Organic Heterostructures by Adina Scott, UWTV 2009
- Connectomics by H. Sebastian Seung, Living Systems DC Salon July 2010
- ~18min35sec "death is the destruction of your connectome" (following a doubts on current cryogenic)
- Body Battery by David Pescovitz, Lab Notes: Research from the Berkeley College of Engineering
- Embedded Technologies: Power From the People, Smithsonian Magazine August 2010
- Anders Sandberg's page on Uploading
- David Chalmers concerns on gradual uploading
- live wikying
- Path:/pub/Events/TeleXLR8-01-ASIM17072010.txt (note that discussions were still going on when I left, thus the log is not complete)
- "Save Transcript..." -> ~/discussion-text.txt
- high-frequency "ack" responses lacking in other interaction means
- discussions with Paola (13/07/2010 and 14/07/2010) on corticogenesis, early spontaneous brain activity, correlation embedded as a physical mechanism, body discovery, ...
- consider doing a short article on security
- Your whole emulated mind (or uploaded mind) will be hijacked and ransomed, here is why and how
- computer security basis
- law basis
- economical basis
- overall basis
- sources for potential solutions
- individual immune system
- group-based immune system (vaccine)
- computer security
- RAM to RAM encrypted network transfer
- weak arguments
- nobody cares about my "instance"
- botnets scan target hosts automatically 24/7 for exploits
- my "instance" would be useless
- if it's precious to anybody (even if not yourself but your entourage) then it has value thus can be worth capturing
- my "instance" will be one in haystack of the Internet, it will be mathematically impossible to find in a realistic time
- just like current P2P swarm, one list (even in a distributed hash-table form) is enough to very efficiently parse an entire network
- it is very unlikely that nobody (including the participant) will make a list of the running "instance", at the early stage that number will probably be very low thus easier to analyze
- see also
- workshop proposed (early August 2010) to Marc for the Paris III Censier-Sorbonne conference
- mindset of "X (mind uploading) is possible, X meets criteria A, B or C (cf above), security is thus an issue" where "X is possible" applies even for those who do not think so
- else those person must refrain from participating
- gedankenexperiment, closure
- consider neuronal growth cones for biotic/abiotic interface and proper neuronal designed circuity
- check http://www.quora.com/Will-human-consciousness-ever-be-transferrable-to-a-computer