Summary

The Dream Machine follows psychologist JCR Licklider and the researchers he brought together to turn his vision into personal computing and the internet. His idea of human-computer symbiosis gave people the goals, judgment, and creative work while computers handled the routine tasks that consumed their time. At ARPA, he used research funding and a talent for connecting people to build a community of fellow believers who could pursue that future beyond his own tenure.

The story follows that community through Project MAC, the ARPANET, and Xerox PARC as interactive computing grew into networked personal machines. Licklider’s lasting contribution was the people and institutions that carried his vision forward, toward a global commons where computers would help people exchange knowledge and think together.

“Boosting mankind’s ability to deal with the complex urgent problems would be an attractive arena in which a younger person might try and make a difference.”
– Douglas Engelbert

“In effect, heuristic reasoning is what allows us to go through life in a chronic state of controlled panic.”

Chappy's Review

A fascinating and engaging book covering the dawn of computers and networking – orbiting around a less-known character in Lick and his vision + community. Great if you’re in need of perspective of the broad strokes of computing + AI history!

Key Takeaways

Notes

Prologue: Tracy's dad

• Came to believe computers would be:
  • Humane, intimate
  • New medium of expression
  • Democratize access to information
  • Foster communities
  • New global commons
• ↑ symbiosis with humans, cohesive whole
• Goal go create a global network of people working on this ↑
  • Goal: connect all their computers into a single network

1: Missouri boys

• Gifted, curious childhood
• Art → engineering → neuroscience
• Brought a unique perspective to the 25y later dawn of computing
• Quiet but natural leader
• Always having fun
• Multi-layered big thinker
• “Always grabbing onto new ideas” / new fields
• Weiner
  • Father was professor, “harshly perfectionist taskmaster” tiger parent
  • College at 11, PhD by 18
  • Worked by intuition, not math + written logic
  • Lost in thought to the point of amnesia
  • Paranoid

2: The last transition

• “Computer” was a pink collar job back in the 30s
• 1939 “desk library” paper helping average people (not institutions) was a radical concept not realized for 40 more years
• WWII radar etc. prioritized over computer development
• Claude Shannon
  • Differential analyzer maintenance
  • “A symbolic analysis of switching circuits” senior thesis “arguably the most influential of the 20th century”
  • Binary arithmetic, logic, programming
• Binary math was the future of computing
• Babbage + Lovelace first computer
• At 40, Von Neumann already only comparable to Einstein
  • Savant at everything
  • “Seemed to have no human failures at all”
• ENIAC project launched 1943
• P/NP
• Turning
  • State machine
  • Basically was the core mechanism for all computers
  • Halting problem
• Feedback is required for systems
  • Biological ↔ computational
  • Defines purpose for living and non-living systems
  • Foundation for AI
• 1943 paper that neural networks are functionally equivalent to Turing machines
• Von Neumann fetch, execute, return
• Stored program invention → public domain
• Shifted us from structure to behavior
• “People could begin to convince as these machines as something fundamentally new. Computing machines had become computers, the devices that implemented software”

3: New kinds of people

• Accepted that the mind was a black box
• Behaviorism vs “mind” unscientific taboo
• Stevens vs Skinner ↑
  • “Bind mysticism vs cold reductionism”
• “The mathematical theory of communication”
• Claude Shannon + information theory 1948
  • 5 parts: information source, transmitter, channel, receiver, destination
  • Abstract clarity
  • Largely kept To himself
• Von Neumann suggested term as information → entropy
  1. Exact same as in physics
  2. People don’t understand it so you’ll always win the argument
  • Finally, someone else suggested “bit” for binary digit and it stuck
• Weiner’s “new science”: Cybernetics
  • Information theory + prediction theory
  • Though stopped after WWII bombs
• ↑ vs Von Neumann’s “calm dispassionate clarity”
• 1949: first book on programming
  • Eg. flow charts, subroutines (now functions or modules)
• Simulation
  • 1954 first use: weather prediction
• Universal constructors + self-replication
  • Eg. DNA
• Weiner viewed VN as “dangerous intellectual hubris” - saber rattling + game theory of nuclear arms race
  • Viewed society as more of a “complex adaptive system”
• Cybernetics + Weiner’s autobiography
  • The world is an organism
  • We have agency
• Information theory “big bang”
  • Conceptual + technical vocabulary, very little antecedence
• Cybernetics book
  • “Impossible for anyone seriously interested in civilization to ignore this book”
  • Information as a fundamental substance - unified communication, computation, control
• The Information Bomb ↑
• Cognitive science: the brain designed to acquire, communicate, and transmute information

4: The freedom to make mistakes

• Forester + Everett MIT
• Up until then, computers had always been conceived of as calculating machines
  • This was constantly monitoring input
  • First realtime computer
• ↑ The Whirlwind Project
• Lick = “top ranked psychologist who thoroughly understood human machine interactions”
  • Own inner quest: how the brain itself works as a system (and in conjunction with computers)
• Focus on core memory
• IBM punch card tabulators → computers
• SAGE computer 1950s
• Cybernetics became too commonplace → died as a movement
  • Failed unification of human + computer info processing science
  • Why? Mostly Weiner falling out of favor + mortality
• Turing
  • Turing test paper mostly ignored at first, now ubiquitous
  • Non-linear system dynamics / chaos theory predicted 10y early
  • Large uncertainty surrounding supposed suicide
• Von Neumann bone cancer
  • “Suffer more when his own mind would no longer function than I have ever seen any human being suffer”
  • Died at 53
• Lick led basically the first cognitive science dept at MIT, 20y early
• Brain’s 7 chunks of working memory
• Noam Chomsky
  • Grammatical machines
  • “Turing machines are the minimum needed for human cognition”
  • BF Skinner book review
• Newel + Simon:
  • Economics + Chicago
  • Bounded rationality
  • Coined satisficing
  • Eg. getting dressed in the morning has billions of possible options
  • Heuristics: ↑ apply to the complex and unexpected
  • “In effect, heuristic reasoning is what allows us to go through life in a chronic state of controlled panic”
  • Theorem proving via computational grammars
• ↑ rules that can generate behavior
  • How the mind can arise from information processing
• George Miller - led the emergence of cognitive psychology

5: The tale of l fig tree and the wasp

• Only uses for computers were calculation and bureaucracies
• Olsen + Anderson starting a transistor computer company
  • 1957 DEC 70k
• Lick epiphany
  • TX2 personal computer
  • Humans best at intuitive processes: perception, goal setting, insight
    • Heuristic mental processes
    • “Carried out by massively parallel networks of neurons operating well below the level of consciousness”
  • ↑ thus, 85% of tasks (time “getting in a position to think”) could be automated and rest of human time could be opened up for “real creativity”
• Left MIT psychology after building a 20y career for computers
• BBN acoustics co
  • Led acoustics → human machine interaction
  • “People liked being collected” - sought out to join elite team?
  • “Hire the most brilliant people you can and stand back”
• Instead of making computers bigger, they’d have to make them more Intuitive
• DEC’s PDP1 delivered to BBN in 1959
• John McCarthy
  • Reinvented interactive computing
  • ”Founded AI”
  • LISP
  • General purpose time sharing (1955)
  • Either monologue or no answers for a long time
  • Tended to come back with answers days later (arrogant)
• McCarthy + Rochester + Minsky + Shannon 6-week conference at Dartmouth on AI sponsored by Rockefeller Foundation
• Linked lists (LISP)
  • Functions
  • Better for bottoms up programming
• ↑ “both data and executable code at the same time”
  • Recursive data theory
• “The problem was everybody had kind of dreamy visions of interacting with equipment, the thing that John did was to spell out the particulars of how you actually go about it.”
• “Collected ideas as avidly as he did people”
• Not leader (all too type A), but integrator + synthesizer + visionary articulator
• 1960 “Man Computer Symbiosis”
  • “It essentially laid out the vision and agenda that would animate US computer research for the next quarter century, and arguably down to the present day”

1. Relationship > tool
2. Real-time command and control
3. Each supplying what they do best

• Predicted by 2010 “intellectually the most creative and exciting in the history of mankind”
• ↑ post hoc “it was just so obvious, but one of us has to go out and write it”
• Programming as a fundamental intellectual skill (functions)
• Models: “any convenient simulation of reality”
  • Models as process > documents
  • Group viewable + manipulatable models
• Biomedical computing
• Computers for everybody

6: The phenomena surrounding computers

• Sputnik moment → “moving science and technology to the center of US policymaking”
• ARPA
• DoD slowly realizing need for “interactive computing”
• ↑ ARPA → Lick “he could reinvent the whole field of computing…”
• Got the $10M CS/AI gov’t grant then went recruiting all around the US
• Douglas Engelbert
  1. “The difficulty of man’s problems is increasing at a greater rate than our ability to cope. We are in trouble”
  2. “Boosting mankind’s ability to deal with the complex urgent problems would be an attractive arena in which a younger person might try and make a difference”
  3. General purpose computer powered info environment (graphical)
  • Dedicated career to “augmenting the human intellect”
• DE wrote manifesto over the next 2y
  • Evolution of symbol manipulation
• “Turn them into a real community”
• Project MAC: machine-aided cognition
• “Programs and data would actually live on the net”
• ↑ “In just a few years, this memorandum to the intergalactic network would become the direct inspiration for the Arpanet which would eventually evolve into today’s internet”
• -—
• Project MAC time share network
• “Information as a utility”
• “The sociology of software”
• Made exchanging software trivial for the first time
• Fault-tolerant systems
  • Eg. firewalls, clocks, backups, passwords
• Newell, Simon, Perlis
• The mouse*
• IBM System 360 → made computers an enterprise standard
  • “The Mythical Man Month”
• Over to IBM - “even a fraction of that marketing power…”
• Lick had…
  • Psych background → saw computers in relation to human brain + creativity
  • Vision + opportunity → guts to go for it
• Taste to select the right projects
• Patience to take the long view: “by creating a community of fellow believers, he guaranteed that his vision would live on after him”
  • “A nationwide movement that had direction, coherence, and purpose”
• University + student focus = “nearly co-opted a substantial portion of the rising generation”
  • Convincing them CS > physics
• Community became a field

7: The intergalactic network

• Nationwide network of all 16 contractors
• Massive timeshare → millions of machines connected as equals
• “Took pride in the notion that someone with a good idea could get $1M within a day”
• Networking details - channels, packets, distributed
• Nationwide packet switching
• Late 60s war activism → gov’t funding sentiment souring
• —
• First 4: UCLA, SRI, Utah, UCSB
• MLS networking + UX showy demo in front of massive audience
• “It would take people time to grasp the significance of something so new”
• “Timeshared interactive computing”
• ”The Woodstock of personal computing” ↑
• BBN won ARPA network contract
• ARPAnet miraculously up in 9mo
• Lick + IBM was a disaster
• Became director of project MAC
• “…again and again took jobs that required him to be an administrator, and then was almost contemptuous of the skills it took to do those jobs right.”
• -—
• Minsky
  • Mission - worked around the clock
  • “Inventing the future”
  • “Why waste your time augmenting human intelligence when humans are basically obsolete?”
• Unix = “Multix without balls”
• Younger generation being driven into cynicism, paranoia, despair by Vietnam
• “On the frontier, man must often chart his course by stars he has never seen”
• Email made the CS community tight knit for the first time
• Packet switching coming of age
• Network working group knew that this was going to change the world
• “Exercising responsibility at a remarkably early age”

8: Living in the future

• Xerox PARC
• Research needs culture + own space to breathe from number crunchers demanding results
• Picked Palo Alto - “it would become the epicenter of computing”
• Silicon Valley coined in 1971
• Bell → Shockley → Fairchild → Intel
• Basic + applied CS labs
• Individual small computers + uncertainty
• Headhunting top university researchers
• Management style: high autonomy - mainly managing personalities
• Goal: “the electronic office”
• “Somehow he’d have to set things up so people would freely follow their own instincts and end up organizing themselves. This is arguably the fundamental dilemma of modern management - not to mention political challenge in any democracy.”
• Get people using it + get feedback → build better technology
• Class 1 → 2 disagreement: anger + shouting → can explain other side’s argument to the other side’s satisfaction
  • Don’t have to believe the other guy, just have to give a fair account of what they’re saying
  • Amazing at 1) unspoken assumptions 2) parroting out facts each side knows
• OOP
• “Personal dynamic medium” vision for a computer
• Right people + right content (man computer symbiosis) + sit back
• Reason for timesharing → personal computer: display requiring majority of processing power
• 1972-32 Alto PC: first personal computer we’d recognize
• Must be distributed personal computing so it’s not a step back
  • Bob Metcalf
• Ether network / Alto Aloha - can send packets into the “ether”
• Connecting ↑: inter-network → internet
• Gateways + Interface 22 spec (written by Robert Khan)
• Need “envelopes” + universal protocol → TCP
• Also needs addressing: internet protocol → TCP/IP
• Smalltalk programming environment
• Software, eg. WYSIWYG → Word
• PARC = “Arpa continued by other means”
• Story about PARC / culture clash + numbers focus de-prioritized in Xerox
• Move to Dallas “was the one clear fuck up”
• Business people talked down to for not knowing about computers (bad)
• ARPA → DARPA, relevance statements, lack of trust, “not like it used to be”
• Lick reappointed, “not like he was” now at 58
• “Creativity is a function of freedom”
• Pragmatic (new) vs intuitive (Lick)
  • Merits to both approaches
• Xerox dealing with geographic + ideological factions
  • Killer demos but rejected by execs dealing with flailing copier business

9: Lick's kids

• Always a sense that he was playing
• MAC (Multi-access computer / machine assisted cognition) → LCS (Laboratory for CS)
• Spiritual + intellectual leader
• Visionary - helped people feel like they were building something bigger
• “The multi-net” - vision was basically the internet but was not a foregone conclusion in 1979
• Information marketplace internet idea
• DEC “mini computer” in 60s + OEMs
• Unix free because AT&T was a regulated monopoly until broken up
• Unix pipes + C
• Bill Joy
• DEC fumbled opportunity for personal computer - didn’t believe in the concept
• Altair - first commercially successful micro computer
• Altair Basic programming language → Microsoft
• 100s of startups building clones
• Late 1970s cultural takeoff
• Homebrew Computer Club + Apple
• BIOS → MSDOS → Microsoft
• Makes people feel like they’re part of the future
• Apple Xerox PARC demo impact mostly exaggerated
  • Gabe energy + focus
  • Epiphany: must be fun to use
• The tragedy is it wasn’t just top-level Xerox failure, but also engineers not knowing how to build / execute for market success (for STAR)
  • Too many features / too slow
  • Closed system
    • Better for enterprise, but couldn’t tap into community thus had to guess right
  • Could’ve just done something simpler
• “If anything, it was an excess of vision at PARC”
• 200+ people thought they had a special bond with Lick at his retirement party
  • “I’m one of Lick’s kids”
• “Actual birth of the internet”: Arpanet switching to TCP/IP in 1983
• NSF ↑ supported in 80s (+Al Gore)
• Forcing regional ISPs to become profitable made internet successful
• NSF net seised to exist in 1995
• WWW for hypertext linking
• Gopher search engine
• “Technology isn’t destiny… the way its capabilities are used is as much a matter of cultural choice and historical accident as politics is or fashion.”
• “He was really the father of it all” - Lick for personal computing and networking

Addendum

Man-Computer Symbiosis (1960)

• In symbiosis, “man will set the goals, determine the criteria, and perform the evaluations. Computer machines will do the routine-izible work that must be done to prepare the way for insights.”
• Estimate: 1980 until AI is good enough for military applications
• Parallel + flexible vs sequential
• Need time sharing + networking “a thinking center”
  • Economics of scale
• Lagging: memory, controls, displays
• Can’t make people learn to type
• ASR

Intergalactic Network memo (1963)

• Need a common open medium
• Better transmission, collaboration of mental models (more powerful than printing press)
• Core of communication: comparing and restructuring mental models
• More important than simply packet switching / routing
• Super-community: all get access to info + resources of others
• Issues: data integrity, software portability
• Estimated cost: $16/h (argued as a steal in comparison to phone calls + travel)
• “Linked consoles” “tele-meetings”
• Oliver: literally AI agent
• Social groups by interest > proximity
• Must make being online a right, not a privilege

The Computer as a Communication Device (1968)

• Need for network communication standards