AI Mind Maintainer

Information about the career pathway for the profession of AI Mind Maintainer. Privacy policy: Third parties advertising here may place and read cookies on your browser; and may use web beacons to collect information as a result of ads displayed here.

Sunday, May 20, 2018

jmpj0520

Slowing down the speed of thought to wait for human input.

The biggest complaint about the JavaScript Artificial Intelligence (JSAI) recently is that the AI output keeps changing faster than the user can reply. Therefore we need to introduce a delay to slow down the AI and let the human user enter a message. In the AudListen() module we insert a line of code to reset the rsvp variable to an arbitrary value of two thousand (2000) whenever a key of input is pressed. In the English-thinking EnThink() module we insert a delay loop to slow the AI Mind down during user input and to speed the AI up in the prolonged absence of user input.

Sunday, May 13, 2018

jmpj0513

Answering of what-think queries with a compound sentence.

We would like our JavaScript Artificial Intelligence (JSAI) to be able to answer queries in the format of "What do you think?" or "What do you know?" We begin in the InStantiate() module by zeroing out the input of a 781=WHAT concept by adding a line of code borrowed from the ghost.pl AI. Then we input "what do kids make" and the AI correctly answers, "KIDS MAKE ROBOTS". However, when we input "what do you think" or "what do you know", the AI does not respond with "I THINK..." or "I KNOW...". Therefore we need to make use of the Indicative() module to generate a compound sentence to be conjoined with the conjunction "THAT". Into the MindBoot() vocabulary we add an entry for the conjunction 310=THAT.

After much trial and error we have gotten the JSAI to respond to the query "what do you think" with "I THINK THAT I AM A PERSON". We let the English-thinking EnThink() module call the Indicative() module first for a main clause with the conjunction "that" and again to generate a subordinate clause. When we ask, "what do i think", the response is "YOU THINK THAT I AM A PERSON". When we inquire "what does god think", the ignorance of the AI engenders the answer "I THINK THAT GOD THINK" which may or may not be a default resort to the ego-concept of self.

Saturday, March 17, 2018

jmpj0317

Restoring JavaScript variable-comments and removing obsolete variables.

Since we assume that many people have made copies of the JavaScript Artificial Intelligence (JSAI) in order to study it, today we carefully make some pressing changes and for each change we provide an explanation by way of justification. First we delete some previously commented-out code-lines which were left in the open-source AI codebase for the sake of continuity, that is, to show that the particular lines of code were on the way out. Thus we remove the commented-out variable "kbyn" from 30jun2011.

Next from the obsolete 20mar07A version of the JSAI we restore comments for some variables and we remove some obsolete variables. We add a link to Consciousness above the Control Panel.

Anyone finding a bug in the AI software may subscribe to the mail-list agi@listbox.com for Artificial General Intelligence (AGI) and report bugs to the AGI community or engage in archived AGI discussion. There is no bug-bounty, other than the glory of the deed.

Friday, March 02, 2018

jmpj0302

Moving the JavaScript AI towards Artificial Consciousness

Two important goals for the AI Mind in JavaScript are the already demonstrated Natural Language Understanding (NLU) and the not-yet-proven Artificial Consciousness. Before we work explicitly on consciousness, we remove the clutter of some obsolete tutorial display code from the MainLoop and elsewhere, so that the program as a whole may be easier to understand and work with.

We have a chance here to demonstrate an entity aware of itself and of some other entity such as a human user conversing with the AI. If we start claiming that our JSAI has consciousness, Netizens will test the AI in various ways, such as asking it a lot of questions. Typical questions to test consciousness would be "who are you" and "who am i". The interrogative pronoun "who" sets the qucon flag to a positive value of one so that the SpreadAct module may activate the necessary concepts for a proper response. We need a way to make the AI concentrate on the subject of any who-query, so that the AI will give evidence of consciousness simply by answering the question.

When we enter "god is person" and then we ask, "who is god", the AI answers "GOD AM A PERSON" -- which sounds wrong but which only requires an improvement in finding the correct form "IS" for the verb "BE".