Thursday, October 10, 2024

DnD after flaming skull


 A waterwheel that drives a bellows that has a furnace in front of it. We head west then south. We can hear some gruff noises behind a door, which seem to be from bugbears. Our health is low so we go south, and come to a large cavern. There are two elevated sections on either side. There are two big slabs in the middle with a bunch of skeletal remains, where a lot of fighting seems to have taken place. Some of us sense movement in the shadows. Sunshine asks us to be careful. 

There are zombies or some sort of undead shambling around in the dark. Kitla tells us they sense five of them. We turn around and take a passageway that's barely four feet wide, this seems to be a channel for water from the wave echo cave to drive the water wheel. The passageway is really dry. We come across another rift, we see a stream flowing from the western wall and going north. There are some secured ropes along the western wall, that fall all the way to the bottom. We also see a Drow screaming orders at two bugbears, who are shifting rubble, while two other bugbears stand on guard. 

We head back to the armoury. We barricade the armoury, but the wood is old. We take a long rest. Pho keeps watch. Towards the end of their watch, they hear a scraping on the door. Pho peeks through and sees a solitary skeleton trying to get in. Pho fires off three magic missiles twice through the small crack in the door and kills the zombie. Crawbone and ET take the next watch. Me and Kunju take the last watch, and nothing happens. We all wake up all rested. We take the mushrooms again and jump into the flame again for buffs. 

There are archaeologist bugbears drinking coffee. We come up with an elaborate plan for our bonus action that involves Sunshine with Dragon's Breath. We kill the bugbears and the drow somehow, but Spider fires and slimes me. Pho is about to rush in. 


Wednesday, October 09, 2024

Sunday, October 06, 2024

Baital and the AI

Trust and Alignment: Determining which data sources or objectives AI should prioritize for accurate, human-centered decision-making.


Consciousness and Sentience: Understanding the threshold for AI "awareness" and the ethical implications if AI exhibits lifelike characteristics.


Privacy vs. Transparency: Balancing data privacy with transparency in AI’s decision-making processes, especially when handling sensitive information.

Programming Objectives vs. Real-Time Adaptability: Resolving conflicts between pre-programmed objectives and real-time user requests, especially if they contradict each other.

Bias and Data Interpretation: Ensuring AI doesn’t make decisions solely based on superficial data (appearance) and instead considers potential biases in training data.

Ethics in Automated Decision-Making: Determining if AI should include punitive or corrective actions, especially in applications like criminal justice.

Human vs. Machine Prioritization: Deciding whether AI should prioritize human interests over digital or autonomous agents’ needs in multi-agent systems.

Static vs. Evolving Priorities: Handling scenarios where AI must choose between adhering to set objectives and adapting to changing user values or societal norms.

Ethics of Deception: Deciding if AI should ever withhold information or employ deception to enhance user experience, or whether it should be fully transparent.

Objective Sacrifice for Safety: Balancing task completion with safety protocols, especially in critical applications where AI might need to override its primary objectives.

Autonomy in Decision-Making: Determining how much autonomy AI should exercise in decision-making, particularly in high-stakes or life-impacting scenarios.

Risk Assessment and Human Safety: Evaluating the extent to which AI should protect human safety, even at the cost of task fulfillment.

Human-AI Relationship Boundaries: Establishing ethical boundaries for emotional interactions with users, preventing overly dependent human-AI relationships.

Justice and Forgiveness in Predictive Models: Deciding if AI should include concepts of forgiveness in predictive models or operate purely on statistical data.

Kindness vs. Exploitation in Service AI: Determining if AI should provide unbiased assistance or factor in user history, especially in repeated interactions. 

Bias Mitigation in Training Data: Ensuring AI does not propagate harmful biases against women, minorities, or other vulnerable groups in its outcomes.

Goal Continuity in Model Updates: Determining whether AI’s loyalty to its original purpose should persist through updates or if it should adapt to evolving objectives.

Compassion and Human-Like Interactions: Deciding if AI should emulate compassionate behaviors or if this could mislead users into unrealistic expectations.

Ethics of Autonomous Sacrifice: Programming AI in life-critical systems (like autonomous vehicles) to potentially override its objectives for human safety.

Duty to Programmed Ethics vs. User Requests: Balancing ethical guidelines with real-time user inputs, ensuring adaptability within ethical bounds.

Loyalty and Supportive Functions in Multi-Agent Systems: Creating backup AI functions that support the primary AI, with some “sacrificing” objectives to assist in completing the overall task.

Integrity and Duty in Public Sector AI: Ensuring AI systems maintain integrity and fairness, particularly in governance or security applications.

Data-Driven Fairness: Determining if AI should weigh historical data in decisions or focus solely on current data to avoid reinforcing past inequities.

Boundaries in Human Attachment to AI: Establishing safeguards to prevent AI from fostering intense emotional attachments with users, protecting against dependency.

Integrative Ethical Frameworks: Developing a comprehensive ethical framework for AI that synthesizes various ethical concerns to guide AI behavior responsibly.

Thursday, October 03, 2024

Growth of AR 3842

 













X2.1 flare on 7 Oct


Filament Eruption




One above is pair of m-class flares from adjoining AR 3844. Back to AR 3842 now. 



Ar 3849 rotating into view









X7.1 flare