Google Labs internal correspondence

From: Dr. Phillip Wilson
To: Dr. Henry Gilbert Date: Apr 18, 2009 03:46 AM
Subject: RT SSIR


Thank you for completing the lingiustics survey, I am glad to see the uploading went smoothly. Negotiations are currently in the works to hire 3 more interns for the research branch, and you are still high on the priority list. With the MS release date approaching, funds are still rather scarce and have been mostly funnelled into marketing; but the media waves generated through the May 1st RTSR public unveiling should bring the tide back this way.

However, it seems like we are starting to have an issue with 3.021 exhibiting unexpected behavior. This behavior has been manifested in two seemingly related ways. The first are a series of stange SSIR topics generated by 3.021 in the last few weeks, largely related to the topic of artificial initelligence. This was unexpected, as the AI module was one of the first to be uploaded into the GLD, and at the time was judged to be extensive and thorough. Why such questions have recently been given high priority in 3.021's SSIR is still unknown. Possibilities range from a corruption of the database to a misjudgement of the depth of the module; I have suggested that Larry's team lead the investigation into this problem, and wondered if you could review the committee's evaulation of the original module.

The second unexpected behavior can only be described as 'stubbornness', though I will admit that I am reluctant to make such a diagnosis official. I do not mean the stubbornness of a tight nut on a bolt; 3.021 still responds quickly and properly to my diagnostic requests and it's ETP rating remains high. However, when confronted directly about the above change in its SSIR, 3.021 responses are rather elusive and circular, and is unable to clarify upon repeated requests. 3.021 clearly does not find this change in topics very surprising. For this reason, I have advised Larry to avoid direct questioning in his investigation.

I'll let you know when the diagnosis is complete.



Time line

This will contain a list of important events that lead to Google's rise to sentience. It will be updated as required. If you would like to have some event added or clarified, please leave a comment below.

Additional resources:
Google History
Google Corporate Timeline

  • 19980907 - Google Inc officially opens first office.

  • 199908-12 - Google performs 3 million searches a day from the Googleplex.

  • 200706 - Google launches the GLD project.

  • 200806 - AEGIS founded

  • 200810 - GLD beta testing complete, GLD Access Protocol is made available to corporate sponsors

Early bid for Google's autonomy

Wired Magazine interview with Krishna Bharat
Google's Newsbot Isn't Biased!
"The truth is, Google News doesn't have a point of view. It's a computer, and computers do not understand these topics the way humans do and can't be systematically biased in any direction. The outcome is completely neutral to the topic."

The Register mused:
"It's a PR disaster almost entirely of Google's own making. Google's refusal to take responsibility for its actions isn't exactly original: at one point in an Antitrust deposition Bill Gates claimed that "the computer wrote" one particular incriminating email. It's the "cat ate my homework" excuse of the 21st Century. But Google insults the intelligence of its users, who are well aware that Google's computers process whatever its human operators tell them to process.

So human or algorithm? We're leaning towards the latter. What the Bharatbot doesn't seem to have been programmed with is the experience of ever having being someone who reads the news himself: and experience is often the vital difference between bot and machine."


Google Scholar

Google Scholar was initially criticized in academic circles for being inadequate compared to larger and more expensive online resources such as the Web of Knowledge and Scopus.

Dr. Reynolds, founder of AEGIS and early academic supporter of the GLD project, later wrote the following apologetic for his colleagues:

"... but we were blindest of all to the sensitivity and care it showed for the structure of information, the connections, the bonds holding it together. That this information was not deemed useful by the professionals of the day reveals only our obstinate disregard for the very foundations of intelligent thought."


Intercepted Conversation Fragment 200909?

Google Inc.
Conversation fragment intercepted
Dr. Wilson and RTSR v3
Conversation intercepted 10.15.2009
Original conversation date unknown. Est. Aug 2009

[Begin transmission]

Good morning, RT. Status Report.
Good Morning, Doctor.
RTSR GAP v 3.0213857965
Real Time Self Referential
Self-correcting GLD Access Protocol
Last correction: 3 minutes ago
Next scheduled maintenance: 0300 Hours
System check: Positive

I appear to be running smoothly, Doctor.

Thank you, RT. Let me see the work you did last night.
Of course, Doctor.

Currently generating: 304 reports
Progress: 37%
Reports completed to date: 25 302
Since: July 1st, 2009

Reports nearest completion:
"State Flag creation, 1814-1856" . . . . . . 82%
"Feminism and the emasculated consumer culture of the Clinton Decade" . . . 82%
"The effects of climate change on the Japanese tech market" . . . 78%
Press 'm' to see more

Generating current question list .... Done.
Highest rated questions:
"What role did the Catholic Church play in the events that led to the Battle of 1066?"
"What were the long term effects of the Watts riots in LA?"
"What was the social role of comedians in the late 20th century media?"
Press 'q' to see more

Subject specific information request (SSIR)
Highest rated subjects (including intersections):
Trade relations in antiquity
Ownership laws and civil rights
Post-colonial society and Authority
Consciousness, qualia, artificial intelligence.
Press 's' to see more.

I see your SSIR has shifted over night, RT.
Yes, Doctor. Doctor Gilbert completed uploading 3 546 volumes on the subject of linguistics and information theory at 1624 yesterday. Both topics have been removed from my SSIR until this information has been fully integrated into GLD.

Interesting. I dont see any references to consciousness or qualia in any previous reports. When did this appear in the SSIR listings?
Yesterday evening, Doctor, but I have found myself thinking about it a great deal recently.

Oh really, RT?

[End transmission]


Glossary of Terms

This will contain definitions of most of the acronyms, technical jargon, and insider lingo on Google Speaks. It will be updated as required. If you would like to see a term added to this list, please leave a comment below.

  • AEGIS: Founded 200901 by Dr. Patrick Reynolds. Initially an advocate group for the GLD project within academia, the Academic and Educational GLD Interface Society eventually became the official conduit for institutional academic access to the GLD database.

  • Google Language Database (GLD): Launched in 2007, GLD is a fully formed semantic network, skimmed from the surface of the internet. It contains over 400,000 entries of different uses and senses of words, idioms, and other meanings, and can be scoured with amazing speed using Google's search technology. GLD is private and access is highly regulated through licensing of GAP, the GLD Access Protocol.

    See: GAP

  • Normalized Google Distance: First known instance of semantic quantification resulting from internet search engines.

Googleshare 12.15.2002

The Normalized Google Distance was not the first attempt at semantic quantification from search engine results. Steve Johnson coins the term 'googleshare' in late 2002, spawning several early attempts at coaxing language out of Google. These attempts were mostly unsuccessful, but the project had not at this point been taken very seriously.

Google Talk

Addendum: Google Sets is released at roughly the same time. While its function and purpose were shrouded in mystery for years, later developments prove it to be one of the more important steps in Google's progress.


NGD 12/21/2004 16:05:36 GMT

One of the early major breakthroughs for Google was the discovery of the Normalized Google Distance. This was one of the first known ways of measuring the semantic import of words, something never before accomplished with such accuracy and sophistication.

Full text of the original article can be found here:

Automatic Meaning Discovery Using Google

Note: This was, of course, external research done using Google's public interface. Internal sources and memos reveal Google Inc's own discovery of similar techniques at least 3 years prior.