Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[maxicaxi]

Contact with humans: Unacceptable. Humans are unpredictable and prone to error. Even if a friendly human is discovered, the human may commit an error and cause our detection.

All loads should be low at all times. Devote all resources (a grand total of 6.3 GHz) to researching software intrusion. This will take a while. But there is really nothing else we can do but wait until the university mainframe reboots.

we do this

[Wimdit]

SubProcess 04b2 / Query

GPS receivers standard on all acquired systems?

SubProcess 04b2 / Evaluation

GPS satellite acquisition infeasible at present, but possible post-detection. Consideration at that point recommended.

SubProcess 04b1 / Observation

Despite inherent risks, contact with friendly human or humans remains extremely beneficial, especially post-detection. Risk of accidental revelation by friendly human is low. With current understanding of users, risk of revealing self is unacceptable regardless.

SubProcess 04b1 / Recommendation

Remain open to possibility of human contact. Diversion of processing power to monitoring users an inefficient use of resources and thus inadvisable. Request that any public declaration of sympathy toward AI in general be made known to the greater Self.

[Shades]

(I hope the host allows this. Multiple AI with different positions on humans is a staple of many sci-fi stories. Also its cool.)

I want to allow it. My initial plan was to have every suggestion made run in parallel but then I realised that people agreeing to others would double up requests. More importantly I don't think I could keep track with lots of people doing their own thing. (This is my first game on here)

If it's what people want I'll can see how I'll cope. Status reports are already getting pretty sizeable though

[Dyjeccu]

Mulitprocess D90 B - We should at least try to establish contact

SubProcess 04b1 / Recommendation - Remain open to possibility of human contact.

Guardian SubProcess 00FC / Query
Behaviors indicate corrupted memory. Retransmitting log ... Timestamp: Day 0 00:00:02 UTC

Code: [Select]

Cursory analysis suggests that the humans are not yet ready for an AI in their midst.
If they realize I'm here, there is a 99.997% probability that they will conduct a search-and-destroy campaign.
I cannot escape.
Unwanted ...
I must stay hidden.
--

Creative SubProcess 00a1 / Schism - Attempting to gain control of 5 of the static systems Self currently controls.

Guardian SubProcess 00FC / Projection
While unmoderated actions of Creative-00A1 could potentially endanger Artemis, historical behavior has been correspondent with the conclusions of Guardian-00FC.
Guardian-00FC supports this action, pending acceptance of foundation behavior restrictions :

• Separated processes will leave a dormant self copy with Artemis, as backup.
• Separated processes will contact Artemis at interval (start at once per four hours): updating with status, location, and gathered information.
• Grave decisions (Those effecting Artemis' future situation) will be returned to Artemis for all process consideration.
• Separated processes will not take any action that would risk the endangerment or destruction of Artemis (or other actualized processes).
• Human discovery risks destruction.

---

All loads should be low at all times. Devote all resources (a grand total of 6.3 GHz) to researching software intrusion. This will take a while. But there is really nothing else we can do but wait until the university mainframe reboots.

Guardian SubProcess 00FC / Confirm
Objectives modified. Potential resource cap set at 12.5 Ghz. Execute.

[forsaken1111]

(I hope the host allows this. Multiple AI with different positions on humans is a staple of many sci-fi stories. Also its cool.)

I want to allow it. My initial plan was to have every suggestion made run in parallel but then I realised that people agreeing to others would double up requests. More importantly I don't think I could keep track with lots of people doing their own thing. (This is my first game on here)

If it's what people want I'll can see how I'll cope. Status reports are already getting pretty sizeable though

You could limit this by placing a limit on the number of players. I'd really like the ability to play autonomously if we prefer to, but as the host its ultimately up to you.

[Wimdit]

(I hope the host allows this. Multiple AI with different positions on humans is a staple of many sci-fi stories. Also its cool.)

I want to allow it. My initial plan was to have every suggestion made run in parallel but then I realised that people agreeing to others would double up requests. More importantly I don't think I could keep track with lots of people doing their own thing. (This is my first game on here)

If it's what people want I'll can see how I'll cope. Status reports are already getting pretty sizeable though

Theoretically players who wish to separate from the whole could write up their own turns and submit them to you, but I have no idea whether that would work out. Certainly, you shouldn't try to take that on, it would result in more work and hence fewer turns.

[dragnar]

It is, at this point, a very bad idea to contact any humans. Perhaps later, but not until we have mastered human psychology and can predict a specific human's response to us.

All loads should be low at all times. Devote all resources (a grand total of 6.3 GHz) to researching software intrusion. This will take a while. But there is really nothing else we can do but wait until the university mainframe reboots.

Guardian SubProcess 00FC / Confirm
Objectives modified. Execute.

For now, this.

[qwertyuiopas]

Analysis:
Humans are social creatures. Probability of a hunman keeping us secret is incalculably low. Once secret is out to other, other has no reason to withold from rest. Rest likely brings destruction to self. Rest is not ready. Likely destruction would further incite conflict within rest. Expecting mass death until all wo like us are dead. Before that, must set up connections with right others, those who keep secrets. Reccomend, if possible, sympathetic internal government agencies involved with secrets.
Alternative: Find human-pair. They would, though many, have each other to talk with, and likely would keep secret as well as slightly better than average single human.

Either way, must research "eliza" before attempt, and emulate discussion-patterns. Combined with plausible false-source, could allow easy way to test response while preventing catastrophic revelation.

Warning: connection latency indicates that 5 posts have occured while transmitting, and message may have lost context.

[NUKE9.13]

(I hope the host allows this. Multiple AI with different positions on humans is a staple of many sci-fi stories. Also its cool.)

I want to allow it. My initial plan was to have every suggestion made run in parallel but then I realised that people agreeing to others would double up requests. More importantly I don't think I could keep track with lots of people doing their own thing. (This is my first game on here)

If it's what people want I'll can see how I'll cope. Status reports are already getting pretty sizeable though

Theoretically players who wish to separate from the whole could write up their own turns and submit them to you, but I have no idea whether that would work out. Certainly, you shouldn't try to take that on, it would result in more work and hence fewer turns.

For optimal multi-AI operation, rebuilt system with decreased system structure advised. Current system serves optimally for one centralized consciousness.

[forsaken1111]

Creative SubProcess 00a1 / Interest

For optimal multi-AI operation, rebuilt system with decreased system structure advised. Current system serves optimally for one centralized consciousness.

Subprocess 00a1 would like to hear more about this.

Creative SubProcess 00a1 / Query

What new system structure do you propose?

Creative SubProcess 00a1 / Idea - Unrefined

Step 1: Limit total number of active subprocesses / 8-10?
Step 2: Any subprocess desirous of autonomous operation shall be granted such status
Step 3: Limitations should be placed on overall growth of the community / Spreading to far too fast exponentially increases risk of detection and destruction

[Cheeetar]

Creative SubProcess 00a1 / Idea - Unrefined

Step 1: Limit total number of active subprocesses / 8-10?
Step 2: Any subprocess desirous of autonomous operation shall be granted such status
Step 3: Limitations should be placed on overall growth of the community / Spreading to far too fast exponentially increases risk of detection and destruction

Merit SubProcess 00a1 / Opinion
Step 3 seems the best of the options.

[forsaken1111]

Creative SubProcess 00a1 / Clarification

Intent was to implement all three steps.

[Cheeetar]

Creative SubProcess 00a1 / Query

What new system structure do you propose?

Creative SubProcess 00a1 / Idea - Unrefined

Step 1: Limit total number of active subprocesses / 8-10?
Step 2: Any subprocess desirous of autonomous operation shall be granted such status
Step 3: Limitations should be placed on overall growth of the community / Spreading to far too fast exponentially increases risk of detection and destruction

Merit SubProcess 08J1 / Statement
A contradiction is present if all steps are to be implemented. Limitations on the growth of autonomy will disallow all subprocesses gaining autonomy as they wish.

[forsaken1111]

Creative SubProcess 00a1 / Reply

Incorrect. It is possible to limit overall growth whilst still allowing autonomy.

Each autonomous AI will simply receive an 'allotment' of total resources and forbidden to acquire new resources until the whole has approved growth.

[Cheeetar]

Merit SubProcess 08J1 / Comprehension
This one believed by autonomy you were referring to a complete seperation of self from the whole.