in 2015 Brett Victor took a picture of his bookshelf -
it's fun to look through
i remember being in awe of an adult's bookshelf, as a child.
it's weird to see it like that.
I also remember this feeling -
I went to the university library
to get a copy of catcher in the rye.
been meaning to read that.
I looked up the number, walked up to the 3rd floor
and found 8 shelves of books -
of people talking about catcher in the rye.
it was weird moment.
is this what the world is?
are the smart people ... really ... like this?
it can't be.
because that is so brutally lame.
books are blogs.
get a physics textbook, flip to a random page,
and see a deep-dive into how Newton's mother was a seamstress, or whatever
even the most prized books are filled
with meandering, indulgent anecdotes.
I actually kind of hate books - you know?
you're trying to learn something
this may be the most precious and sincere act -
and you open a book and just get dragged through endless tangents some author found amusing.
people talk about the smell of books,
and we very politely don't laugh at them.
do you think shakespeare wanted you to smell his play.
In some ways,
being romantic about books is a form of anti-intellectualism.
we should care about the information.
if you know a thing, and you read about it in a book again,
then what are you doing is not healthy.
current size of wikipedia
I got involved in wikipedia as a early -
I was 16, and it made perfect sense.
it was obvious that it was working.
people who were around then -
should not forget how brutally the project was mocked.
and the predictable people who did the mocking.
i'm 36 now, and the opposite is happening.
wikipedia is failing, and nobody has noticed.
The idea in wikipedia is that
all information is editable text -
It's not a database.
Here's the markup for the first sentence of the Albert Einstein wikipedia article:
its information is buried in an ad-hoc, and mostly un-parsable text format.
yes, arabic editors must write it in right-to-left, somehow:
It's hard to describe how much of a serious problem this is.
wikipedia's information can't be used -
it's wild, actually.
there's also categories -
when someone in Category:Blonde Cheerleaders
dies their hair, what then?
Category:Formerly Blonde Cheerleaders?
take a trip to today's Categories for deletion -
see people debating Category:Goth Musicians etc.
it can make you weep.
and also rot:
which is what they call it when you change something
and unknowingly cascade problems elsewhere.
there are 8,550,441 redirects, for example.
these are created by hand, every day.
they are a vast array of typos and combinations of title-case.
when a page is renamed, its redirects and links must somehow change.
there are also 634,755 different kinds of templates -
used scattered throughout.
templates nest within eachother, and can't be edited,
so editors just duplicate them:
what should we use for horse ages?
Category:People
→ Category:Musicians
→ Category:Singers
→ Category:American Idol
→ Category:Books about American Idol
wikipedia worked when all its graphs were going up,
but now most are going down.
the problems with the project are obscene.
Danny Hillis likes to say in the 80's he was heckled at a conference-
the guy joked about putting computers in doorknobs
then when he returned a decade later,
the hotel had RFID in every door knob.
people listen to him now
There's this nagging idea that there could be
a universal data format.
- not one individual format -
but one that's self-documenting
so can be re-used in unplanned ways.
There's also this nagging idea
that we could describe knowledge
in a systematic way
finally do it right.
like a big pyramid, or a big dictionary,
or a clever database.
... or some structure.
Douglas Lenat tried it -
he's my favourite guy.
his project spiraled. They all did.
it seems like a good thing to do -
but what then?
when is it finally working.
because it's never finally working.
in some ways, we're better at creating these sprawling projects today -
crowdsourcing is the (unfortunate) term,
for strangers genuinely collaborating
productively on the internet.
in 2004, after google's IPO, as wikipedia was booming,
- before the iPhone -
a crowdsourced knowledge database was completely obvious.
Danny Hillis raised a mountain of easy money,
he hired a whole lab at MIT.
made a deal to scan the entire Stanford library
they bought a gigantic office on Mission st.
it was happening.
That winter, we bought my grandmother a new laptop.
she opened her browser, and typed in 'swine flu'.
Bing displayed data from the infections disease schema that I wrote.
stupidly, i said 'hey i made that!'
my grandma was like - good job sweetie.
freebase had a lot of problems that wikipedia did not.
on, wikipedia you just type into a box -
freebase became an interface problem -
how can users make a graph?
do they understand?
they went around in circles.
it felt like a gross corporate thing, and it was.
google bought and killed it in 2010.
it's now a $100 million dollar CSV file,
and one of the most depressing anecdotes of software history.
and we're left with wikipedia
this is the infurating part -
and I start sweating, when I think about it -
if you care about what's in books,
then you enter software development
or a software-adjascent life path.
how could you not?
we have all that knowledge there, and we can't read it.
... are you just gonna read a bunch?
and call that a career?
is that really the plan.
there's this joke in computer-science:
'oh you're in comp-sci,
cool!
can you help me setup my printer?'
this is really the situation.
- we can reform human knowledge,
and we haven't gotten around to that yet.
every prick with some app -
the young people productivity hacking.
we're not disrupting anything.
we're just becoming middle-class.
most people buy a house.
we could scan the books.
we could combine them all into one good book.
it would be better then 10 million shit ones.
Everyone from ancient greece, and the middle ages, and revolutionary france - they would all agree.
They would say - you guys can do that??
and you just ... didn't?
.. what the fuck.
and forgive me-
isn't this the whole idea of a library? of an education?
to find and honour the good ideas.
to pick the right ones
and not loose them
it's a very sweet and good impulse,
which i really thought we fostered.
Have you ever looked at a cats paw?
They have knives in their hands.
birds do not think, at all.
Dogs have a whole section of their face with fangs on it.
we are librarians
we are the literary, nerd species.
We are not the tough guys, man.
that's it, for us. building up the books.
we do that.
i worry that our
good librarians
are very lost, today.
On the internet, we're talking over eachother,
all day and then the next day.
it's 90,000 shelves about catcher in the rye.
Its time to stop doing this.
In this way,
wikipedia really is ... it.
It's not some shit website.
We build it well, or we don't.
we are not succeeding,
wikipedia is failing in front of us,
and the consequences of this are really dire.
I mean, all the wires are in place.
typing is fun -
words weigh and cost nothing.
labour is free and endless.
our computers can do a trillion things a second.
we are simply not cooperating.
we are not building something helpful.
today smart people just read.
until their eyes hurt
and then just make more text.
we're stuck.
