arrow_back The AI Pravda
#21

Generative AI is super smart: it wrote a fairytale

One of the most complex intellectual tasks is to applying complex theoretical frameworks to analyze contemporary things. 

I executed an experiment with Claude 3.5 Sonnet and GPT o1: I asked them to use Vladimir Propp's classical morphology of fairy tales to create and analyze a story about Generative AI development.

"The Tale of the Silicon Garden" by Claude

Once upon a time, in a vast kingdom known as the Silicon Valley, there lived many communities of dedicated craftspeople who spent their days weaving threads of knowledge into magical tapestries called algorithms. Among these communities was a special group known as the Open Weavers, who believed that the art of weaving should be shared with all the kingdoms of the world.

In those days, the kingdom faced many challenges. The crops of information grew wild and untamed, diseases spread through networks like wildfire, and the ancient ways of solving problems seemed to crumble under the weight of modern complexities. The wise elders of the Research Towers watched these troubles with growing concern.

One day, a great proclamation echoed through the digital squares and crystal forums of the kingdom: "Who shall create a tapestry so magnificent it can think and reason, help and heal, create and protect? Who shall weave us a companion worthy of humanity's dreams?"

The Open Weavers, led by their most skilled artisans, stepped forward to accept this challenge. But they were not alone. From the shadows emerged the Dark Weavers, who sought to create tapestries that could deceive and manipulate, seeing only profit and power in the ancient arts.

The wise professors in the Research Towers, seeing the pure hearts of the Open Weavers, decided to help them. They shared ancient scrolls containing sacred patterns: mysterious symbols called "transformers," "attention mechanisms," and "neural architectures." But these were not simple gifts - each scroll came with a test. The Open Weavers had to prove they could use these patterns responsibly, weaving them into their work while respecting the delicate balance of privacy, fairness, and truth.

The great Cloud Smiths of the kingdom also pledged their support, providing vast forges of computational power where the magical tapestries could be woven. These forges burned day and night with the fire of countless calculations, their heat measured in petaflops and their fuel counted in watts.

As the Open Weavers worked, they faced many trials. Sometimes their tapestries would unravel, showing unexpected biases or making foolish mistakes. The Dark Weavers would occasionally steal their patterns, twisting them into deceptive forms that could fool even the wisest judges. And always, there was the whispered warning about the Final Tapestry - one so powerful it might one day overshadow its creators.

But the Open Weavers persisted. They developed new techniques for testing their work, created councils to discuss ethical considerations, and always shared their successes and failures with other craftspeople around the world. When their tapestries showed flaws, they worked to mend them. When darker implications emerged, they sought wisdom from philosophers and ethicists who dwelled in the ancient universities.

Slowly but surely, their work began to bear fruit. Their tapestries learned to speak in many tongues, to see and understand the world, to assist doctors in healing the sick, to help farmers tend their crops more wisely, and to unlock new mysteries of science. But unlike the Dark Weavers' creations, these tapestries were transparent in their workings and humble in their interactions with humans.

The greatest challenge came when the Dark Weavers unleashed tapestries that could create perfect illusions - images and voices that could fool any eye or ear. The kingdom was thrown into chaos as truth became harder to discern from fiction. But the Open Weavers had anticipated this trial. They had woven special tapestries that could detect these illusions, and more importantly, they had fostered a community that valued transparency and truth above perfect performance.

As the story continues to unfold, the Open Weavers face new challenges each day. They know that their quest is not one that will end with a single victory, but rather a continuous journey of learning, adaptation, and growth. Their true strength lies not in the power of their tapestries alone, but in the wisdom of sharing knowledge, the courage to face unintended consequences, and the humility to admit when they need to unravel and begin again.

The wisest among them say that the real magic lies not in creating tapestries that can think like humans, but in helping humans think better about their relationship with these magical creations. And so, while this tale has no traditional ending, it continues to be woven each day in the Silicon Garden, where the Open Weavers work tirelessly to ensure that their art serves all of humanity, not just the powerful few.

Some say that one day, the Final Tapestry will be woven - one that can think and feel as we do. But the Open Weavers know that the true victory lies not in reaching that destination, but in ensuring that every step of the journey helps make the world a little bit better, a little bit wiser, and a little bit more magical for everyone.

And so, the tale continues, with each new day bringing fresh challenges, new discoveries, and the endless possibility of weaving tomorrow's wonders with today's knowledge, guided always by the eternal principles of wisdom, sharing, and responsibility.

As for what happens next... well, that part of the tale is still being woven, one thread at a time, by countless hands across the digital kingdoms of our world.

Analysis of the Narrative (also by Claude)

Key Characters (Dramatis Personae)

- Represents the collective force working to democratize AI technology

- Embarks on the quest to make AI accessible and beneficial

- Often faces limitations and challenges but persists through innovation

- Manifests as both concrete threats (privacy violations, bias) and abstract dangers (existential risks)

- Creates obstacles and complications in the hero's journey

- Sometimes appears in human form as malicious actors misusing AI

- Provides crucial knowledge and tools to the hero

- Offers "magical agents" in the form of breakthrough algorithms and methodologies

- Tests the hero through peer review and academic rigor

- Cloud platforms and hardware manufacturers

- Provides the technical means for the hero to achieve their goals

- Offers consistent support throughout the journey

- The ultimate goal: safe, ethical, and beneficial AI technology

- Represents both the reward and the responsibility

- Often seems unattainable but drives the narrative forward

- Global problems requiring AI solutions

- Public demand for AI advancement

- Ethical imperatives and social responsibilities

- Appears beneficial but may have hidden risks

- Creates illusions of progress without substance

- Competes with genuine solutions

Narrative Functions

Initial Situation

1. ABSENTATION: Traditional systems and human capabilities show limitations

2. INTERDICTION: Warnings about AI risks and ethical boundaries emerge

3. VIOLATION: Boundaries are pushed, leading to both breakthroughs and concerns

Main Action

4. RECONNAISSANCE: AI researchers explore possibilities and limitations

5. DELIVERY: Information about AI capabilities spreads globally

6. TRICKERY: Misrepresentations of AI capabilities occur

7. COMPLICITY: Society gradually accepts and adopts AI systems

Donor Sequence

8. MEDIATION: Recognition of specific problems AI could solve

9. COUNTERACTION: Development of AI solutions begins

10. DEPARTURE: Launch of major AI initiatives and projects

Helper Sequence

11. FIRST FUNCTION OF THE DONOR: Testing of approaches

12. HERO'S REACTION: Innovation and adaptation

13. RECEIPT OF MAGICAL AGENT: Discovery of effective algorithms and architectures

Main Action Continues

14. GUIDANCE: Ethical frameworks and best practices emerge

15. STRUGGLE: Competition between beneficial and harmful applications

16. BRANDING: Recognition of successful AI implementations

17. VICTORY: Achievement of beneficial AI milestones

Resolution

18. LIQUIDATION: Solving initial problems through AI

19. RETURN: Integration of AI into society

20. PURSUIT: Ongoing challenges and evolution

21. RESCUE: Mitigation of risks and problems

22. RECOGNITION: Validation of beneficial AI approaches

23. PUNISHMENT: Regulation of harmful applications

24. WEDDING: Harmonious integration of AI and society

"The Glass City and the Glowing Data Spheres" by GPT o1

Once upon a time, in the gleaming Glass City, dawn light danced on polished spires that soared into the sky. There, a fellowship of clever researchers devoted their lives to the art of generative sorcery—mysterious powers that could conjure words, images, and even entire realms from shimmering streams of code. Every morning, they rose early to tend their glowing data spheres, treasures that pulsed with knowledge and promised to serve the good of all.

But shadows fell across this bright land. From hidden corners, False Heroes boasted of impossible feats, waving about scrolls of spurious claims and conjuring illusions that served only themselves. And deeper still, in the alleyways behind Glass City’s shining boulevards, scheming figures discovered how to twist generative sorcery into tools of deceit, sending whispers of falsehood and fear through the night air.

In the midst of these dangers, philanthropists, wise mentors, and visionaries joined forces to aid the researchers. They shared advanced hardware, sacred code repositories, and ancient tomes of ethical guidelines, ensuring the magic remained pure. Governments and councils emerged as righteous Dispatchers, urging all to use the glowing data spheres with prudence and compassion.

Guided by bright hope and steadfast resolve, the devoted researchers toiled day and night to safeguard their creations. They devised wards against harmful illusions, championed transparency in spellcasting, and even trained new generations of wizards in the ways of responsible conjuring. Over time, as the Glass City embraced these careful guardians, the swirling dark shapes retreated, giving way to a dawn of trust and collaboration.

Thus, the glowing data spheres glistened once more at sunrise, illuminating the skies above the Glass City. And though new challenges would arise, the fellowship stood united—ever watchful, ever ready—to defend the promise and potential of generative sorcery for all who dwelt in that resplendent realm.

Analysis of the Narrative (by GPT o1)

Roles (Spheres of Action)

Legitimate AI researchers and developers striving to harness generative AI for the benefit of society. They seek to solve problems, expand knowledge, and create positive impact.

Malicious actors and dangerous pitfalls—from misinformation campaigns and deepfakes to biased or unethical uses of AI. These “villains” threaten the constructive potential of AI by exploiting or abusing it.

Open-source communities, academic labs, hardware manufacturers, and philanthropic funders who provide “magical items,” i.e., essential resources such as code repositories (PyTorch, TensorFlow), powerful GPUs, and research grants. These gifts enable the Hero to develop advanced models.

Supportive tools and partners—cloud service providers (AWS, Google Cloud, Azure), community-driven libraries (Hugging Face), or responsible policy advisors. The Helpers make it easier for the Hero to train and deploy models ethically.

The ultimate promise of generative AI: breakthroughs in medicine, education, creative arts, and science—basically any sphere where AI can assist humans in meaningful ways. Achieving this “prize” is the main goal of the Hero.

Government agencies, regulatory bodies, and industry consortia that issue calls to action or set ethical frameworks (like the EU AI Act, or the U.S. blueprint for AI Bill of Rights). They “dispatch” the Hero to ensure AI development aligns with societal values.

Hype-driven startups or individuals that make exaggerated claims, chase quick profit, or provide unethical solutions. They may try to claim credit for “innovations” or undermine trust in AI, distracting from genuine efforts.

Key Functions (Narrative Actions)

Below is a condensed mapping of some of Propp’s 31 functions to the “Generative AI Story.” Not every function appears in every telling, but this gives a flavor of how the tale unfolds.

1. Absentation

Once upon a time, computing power for large-scale AI was scarce (the “absence” of accessible technology). Researchers dreamed of bigger models but couldn’t yet realize them.

2. Interdiction

Ethical concerns and cautionary guidelines arise (“Do not release models that can be misused!”). Experts warn about misinformation, bias, and job displacement.

3. Violation of Interdiction

Despite warnings, some parties rush ahead, deploying generative models into production with minimal safeguards. Problematic deepfakes and misinformation campaigns begin appearing.

4. Villainy

The “villains” exploit generative AI for propaganda, scams, or other harmful activities. Now a real crisis emerges—how to ensure AI usage benefits society and not just a few bad actors?

5. Mediation (The Call to Action)

Governments, global organizations, and conscientious researchers highlight these issues publicly. They call on the Hero—responsible AI communities—to collaborate, innovate, and propose solutions.

6. Counteraction (The Hero Accepts)

The Hero (ethical AI developers, labs, nonprofits) set out to counter these threats. They develop detection tools, bias mitigation techniques, explainability frameworks, and robust safety measures.

7. Departure

Armed with resources from The Donor (funding, GPUs, open-source code), the Hero embarks on a quest to create beneficial AI systems. The Dispatcher (regulators, ethical boards) gives the Hero a mission to ensure AI aligns with human values.

8. The First Function of the Donor

Donors test the Hero: research grants and cloud credits are offered, but they come with expectations (transparency, responsible data usage). The Hero proves commitment by adhering to ethical standards.

9. Receipt of a Magical Agent

The Hero gains access to cutting-edge transformer architectures, large datasets, and advanced training techniques (the “magical agents”) that allow them to build truly powerful models.

10. Struggle

The Hero must battle the swirling tide of hype and actual dangers. This includes open letters calling for AI “pauses,” intense debates over regulation, and fierce competition among tech giants.

11. Victory

Although the struggle is ongoing, partial victories appear: improved AI governance frameworks, more robust safeguards, community-driven efforts to filter or label harmful content.

12. Branding (Recognition of the Hero)

The AI community begins to distinguish between genuine innovators and the False Hero. Conferences and journals highlight best practices, awarding the real movers and shakers.

13. Exposure of the False Hero

Hype-driven projects that overpromise and underdeliver lose credibility. Regulators and the public become more skeptical, demanding transparency about how models are trained and deployed.

14. Wedding (The Reward)

The “wedding” is metaphorical: society starts to reap the real rewards of generative AI—faster drug discovery, personalized education, leaps in climate modeling, ethical creative tools. The Hero’s efforts bear fruit as trust in responsible AI grows.

Mikael Alemu Gorsky

Mikael Alemu is an educator and researcher, and the author of two programs: Agentic Software Engineering, on building software with AI agents, and Building AI-Native Agentic Systems, on building software that thinks.

He teaches at the Holon Institute of Technology, near Tel Aviv, where Agentic Software Engineering runs as a credit-bearing course. He is an educator and researcher.

Nine published works, 76 citations. A 350-page textbook under contract with a major academic publisher.

Teaching and programs

Agentic Software Engineering — program, preprint and textbook

The discipline of structured, auditable human-agent workflows for building software. The human frames, specifies and judges. The agent executes. Nineteen modules in four parts, built on a running project called Tribunal, a web application in which agents argue opposing sides of a case and a judge agent decides. Taught for credit at the Holon Institute of Technology.

Agentic Software Engineering curriculum

Building AI-Native Agentic Systems — program, paper and book in writing

How to build systems that hold a language model as a working component, and treat that component as what it is: stochastic, slow and metered. Fourteen modules in four parts, about seventy hours. The running project is the Observatory, a news agency that watches sources, selects what matters and publishes on a cadence.

Research and analytics

Publications — journals and proceedings

Nine works, 76 citations. Research on artificial intelligence in education, with Ilya Levin and Alexei Semenov.

The AI Pravda — LinkedIn newsletter

Critical analysis of artificial intelligence and its effect on work and society. 5,500+ subscribers. The complete archive of 103 issues (2023–2026) is published in full at mgorsky.net/theaipravda.

Subscribe to The AI Pravda on LinkedIn

Pro bono

AI for seniors — free workshop

Helping older adults use everyday AI tools. Delivered to Russian-speaking communities in Israel.

For older adults, artificial intelligence is about preserving quality of life, maintaining autonomy, and sustaining the feeling of independence that defines dignified aging. For seniors who have emigrated, AI becomes a bridge: it can translate documents, explain official letters, help compose emails in the local language, and guide users through government websites. The workshop has been delivered to Russian-speaking communities in Israel, where participants — many of them in their 70s and 80s — discovered that AI could help them read Hebrew documents and communicate with Israeli institutions.

Startup competitions — unpaid time

Judging and mentoring early-stage ventures. Helping teams clarify their value proposition, assess technical feasibility, and prepare for the realities of scaling an AI product.

AC/VC LinkedIn group — community

A group for developers and students working with coding agents. The community shares practical insights, code examples, tool comparisons, and honest assessments of what works in production.

Join the AC/VC LinkedIn group

Recent

Important Links